id sid tid token lemma pos fcis-29036 1 1 frontiers frontier NOUN fcis-29036 1 2 in in ADP fcis-29036 1 3 computing computing NOUN fcis-29036 1 4 and and CCONJ fcis-29036 1 5 intelligent intelligent ADJ fcis-29036 1 6 systems system NOUN fcis-29036 1 7 issn issn VERB fcis-29036 1 8 : : PUNCT fcis-29036 1 9 2832 2832 NUM fcis-29036 1 10 - - SYM fcis-29036 1 11 6024 6024 NUM fcis-29036 1 12 | | NOUN fcis-29036 1 13 vol vol NOUN fcis-29036 1 14 . . PROPN fcis-29036 2 1 11 11 NUM fcis-29036 2 2 , , PUNCT fcis-29036 2 3 no no INTJ fcis-29036 2 4 . . NOUN fcis-29036 2 5 1 1 NUM fcis-29036 2 6 , , PUNCT fcis-29036 2 7 2025 2025 NUM fcis-29036 2 8 5 5 NUM fcis-29036 2 9 design design NOUN fcis-29036 2 10 and and CCONJ fcis-29036 2 11 implementation implementation NOUN fcis-29036 2 12 of of ADP fcis-29036 2 13 an an DET fcis-29036 2 14 automatic automatic ADJ fcis-29036 2 15 white white ADJ fcis-29036 2 16 sugar sugar NOUN fcis-29036 2 17 impurity impurity NOUN fcis-29036 2 18 detection detection NOUN fcis-29036 2 19 system system NOUN fcis-29036 2 20 based base VERB fcis-29036 2 21 on on ADP fcis-29036 2 22 multispectral multispectral ADJ fcis-29036 2 23 imaging imaging NOUN fcis-29036 2 24 xin xin PROPN fcis-29036 2 25 zhong zhong PROPN fcis-29036 2 26 1 1 NUM fcis-29036 2 27 , , PUNCT fcis-29036 2 28 li li PROPN fcis-29036 2 29 luo luo PROPN fcis-29036 2 30 1 1 NUM fcis-29036 2 31 , , PUNCT fcis-29036 2 32 * * PROPN fcis-29036 2 33 , , PUNCT fcis-29036 2 34 xiangyang xiangyang PROPN fcis-29036 2 35 yu yu PROPN fcis-29036 2 36 2 2 NUM fcis-29036 2 37 , , PUNCT fcis-29036 2 38 yefan yefan X fcis-29036 2 39 cai cai X fcis-29036 2 40 3 3 NUM fcis-29036 2 41 , , PUNCT fcis-29036 2 42 weibing weibe VERB fcis-29036 2 43 hong hong NOUN fcis-29036 2 44 3 3 NUM fcis-29036 2 45 1 1 NUM fcis-29036 2 46 school school NOUN fcis-29036 2 47 of of ADP fcis-29036 2 48 physics physics PROPN fcis-29036 2 49 & & CCONJ fcis-29036 2 50 optoelectronic optoelectronic ADJ fcis-29036 2 51 engineering engineering NOUN fcis-29036 2 52 , , PUNCT fcis-29036 2 53 guangdong guangdong PROPN fcis-29036 2 54 university university PROPN fcis-29036 2 55 of of ADP fcis-29036 2 56 technology technology PROPN fcis-29036 2 57 , , PUNCT fcis-29036 2 58 guangzhou guangzhou PROPN fcis-29036 2 59 guangdong guangdong PROPN fcis-29036 2 60 , , PUNCT fcis-29036 2 61 510000 510000 NUM fcis-29036 2 62 , , PUNCT fcis-29036 2 63 china china PROPN fcis-29036 2 64 2 2 NUM fcis-29036 2 65 school school NOUN fcis-29036 2 66 of of ADP fcis-29036 2 67 physics physics PROPN fcis-29036 2 68 , , PUNCT fcis-29036 2 69 sun sun PROPN fcis-29036 2 70 yat yat PROPN fcis-29036 2 71 - - PUNCT fcis-29036 2 72 sen sen PROPN fcis-29036 2 73 university university PROPN fcis-29036 2 74 , , PUNCT fcis-29036 2 75 state state NOUN fcis-29036 2 76 key key ADJ fcis-29036 2 77 laboratory laboratory NOUN fcis-29036 2 78 of of ADP fcis-29036 2 79 optoelectronic optoelectronic ADJ fcis-29036 2 80 materials material NOUN fcis-29036 2 81 and and CCONJ fcis-29036 2 82 technologies technology NOUN fcis-29036 2 83 , , PUNCT fcis-29036 2 84 nanchang nanchang PROPN fcis-29036 2 85 research research PROPN fcis-29036 2 86 institute institute PROPN fcis-29036 2 87 , , PUNCT fcis-29036 2 88 guangzhou guangzhou PROPN fcis-29036 2 89 guangdong guangdong PROPN fcis-29036 2 90 , , PUNCT fcis-29036 2 91 510000 510000 NUM fcis-29036 2 92 , , PUNCT fcis-29036 2 93 china china PROPN fcis-29036 2 94 3 3 NUM fcis-29036 2 95 guangzhou guangzhou PROPN fcis-29036 2 96 guangxin guangxin PROPN fcis-29036 2 97 technology technology PROPN fcis-29036 2 98 co. co. PROPN fcis-29036 2 99 , , PUNCT fcis-29036 2 100 ltd ltd PROPN fcis-29036 2 101 . . PROPN fcis-29036 2 102 , , PUNCT fcis-29036 2 103 guangzhou guangzhou PROPN fcis-29036 2 104 guangdong guangdong PROPN fcis-29036 2 105 , , PUNCT fcis-29036 2 106 510000 510000 NUM fcis-29036 2 107 , , PUNCT fcis-29036 2 108 china china PROPN fcis-29036 2 109 * * PUNCT fcis-29036 2 110 corresponding correspond VERB fcis-29036 2 111 author author NOUN fcis-29036 2 112 : : PUNCT fcis-29036 2 113 li li PROPN fcis-29036 2 114 luo luo PROPN fcis-29036 2 115 abstract abstract PROPN fcis-29036 2 116 : : PUNCT fcis-29036 2 117 traditional traditional ADJ fcis-29036 2 118 methods method NOUN fcis-29036 2 119 for for ADP fcis-29036 2 120 removing remove VERB fcis-29036 2 121 impurities impurity NOUN fcis-29036 2 122 from from ADP fcis-29036 2 123 sugar sugar NOUN fcis-29036 2 124 are be AUX fcis-29036 2 125 not not PART fcis-29036 2 126 very very ADV fcis-29036 2 127 inefficient inefficient ADJ fcis-29036 2 128 and and CCONJ fcis-29036 2 129 have have VERB fcis-29036 2 130 limited limit VERB fcis-29036 2 131 accuracy accuracy NOUN fcis-29036 2 132 . . PUNCT fcis-29036 3 1 this this DET fcis-29036 3 2 study study NOUN fcis-29036 3 3 designed design VERB fcis-29036 3 4 and and CCONJ fcis-29036 3 5 implemented implement VERB fcis-29036 3 6 an an DET fcis-29036 3 7 automatic automatic ADJ fcis-29036 3 8 white white ADJ fcis-29036 3 9 sugar sugar NOUN fcis-29036 3 10 impurity impurity NOUN fcis-29036 3 11 detection detection NOUN fcis-29036 3 12 system system NOUN fcis-29036 3 13 based base VERB fcis-29036 3 14 on on ADP fcis-29036 3 15 multi multi ADJ fcis-29036 3 16 - - ADJ fcis-29036 3 17 modal modal ADJ fcis-29036 3 18 image image NOUN fcis-29036 3 19 fusion fusion NOUN fcis-29036 3 20 technology technology NOUN fcis-29036 3 21 to to PART fcis-29036 3 22 detect detect VERB fcis-29036 3 23 impurities impurity NOUN fcis-29036 3 24 in in ADP fcis-29036 3 25 white white ADJ fcis-29036 3 26 sugar sugar NOUN fcis-29036 3 27 efficiently efficiently ADV fcis-29036 3 28 and and CCONJ fcis-29036 3 29 accurately accurately ADV fcis-29036 3 30 . . PUNCT fcis-29036 4 1 the the DET fcis-29036 4 2 system system NOUN fcis-29036 4 3 utilizes utilize VERB fcis-29036 4 4 4 4 NUM fcis-29036 4 5 multi multi ADJ fcis-29036 4 6 - - ADJ fcis-29036 4 7 spectral spectral ADJ fcis-29036 4 8 cameras camera NOUN fcis-29036 4 9 and and CCONJ fcis-29036 4 10 an an DET fcis-29036 4 11 rgb rgb PROPN fcis-29036 4 12 camera camera NOUN fcis-29036 4 13 to to PART fcis-29036 4 14 acquire acquire VERB fcis-29036 4 15 high high ADJ fcis-29036 4 16 - - PUNCT fcis-29036 4 17 resolution resolution NOUN fcis-29036 4 18 images image NOUN fcis-29036 4 19 of of ADP fcis-29036 4 20 the the DET fcis-29036 4 21 samples sample NOUN fcis-29036 4 22 . . PUNCT fcis-29036 5 1 by by ADP fcis-29036 5 2 extending extend VERB fcis-29036 5 3 the the DET fcis-29036 5 4 input input NOUN fcis-29036 5 5 channels channel NOUN fcis-29036 5 6 of of ADP fcis-29036 5 7 the the DET fcis-29036 5 8 attention attention NOUN fcis-29036 5 9 u u ADJ fcis-29036 5 10 - - ADJ fcis-29036 5 11 net net ADJ fcis-29036 5 12 model model NOUN fcis-29036 5 13 to to ADP fcis-29036 5 14 thirteen thirteen NUM fcis-29036 5 15 , , PUNCT fcis-29036 5 16 the the DET fcis-29036 5 17 recognition recognition NOUN fcis-29036 5 18 rate rate NOUN fcis-29036 5 19 of of ADP fcis-29036 5 20 impurity impurity NOUN fcis-29036 5 21 categories category NOUN fcis-29036 5 22 is be AUX fcis-29036 5 23 improved improve VERB fcis-29036 5 24 by by ADP fcis-29036 5 25 approximately approximately ADV fcis-29036 5 26 60 60 NUM fcis-29036 5 27 % % NOUN fcis-29036 5 28 , , PUNCT fcis-29036 5 29 significantly significantly ADV fcis-29036 5 30 enhancing enhance VERB fcis-29036 5 31 the the DET fcis-29036 5 32 detection detection NOUN fcis-29036 5 33 performance performance NOUN fcis-29036 5 34 of of ADP fcis-29036 5 35 the the DET fcis-29036 5 36 model model NOUN fcis-29036 5 37 . . PUNCT fcis-29036 6 1 the the DET fcis-29036 6 2 experimental experimental ADJ fcis-29036 6 3 results result NOUN fcis-29036 6 4 show show VERB fcis-29036 6 5 that that SCONJ fcis-29036 6 6 the the DET fcis-29036 6 7 system system NOUN fcis-29036 6 8 performs perform VERB fcis-29036 6 9 excellently excellently ADV fcis-29036 6 10 in in ADP fcis-29036 6 11 the the DET fcis-29036 6 12 image image NOUN fcis-29036 6 13 segmentation segmentation NOUN fcis-29036 6 14 task task NOUN fcis-29036 6 15 , , PUNCT fcis-29036 6 16 achieving achieve VERB fcis-29036 6 17 an an DET fcis-29036 6 18 average average ADJ fcis-29036 6 19 intersection intersection NOUN fcis-29036 6 20 over over ADP fcis-29036 6 21 union union NOUN fcis-29036 6 22 ( ( PUNCT fcis-29036 6 23 miou miou NOUN fcis-29036 6 24 ) ) PUNCT fcis-29036 6 25 of of ADP fcis-29036 6 26 92.70 92.70 NUM fcis-29036 6 27 % % NOUN fcis-29036 6 28 and and CCONJ fcis-29036 6 29 an an DET fcis-29036 6 30 average average ADJ fcis-29036 6 31 pixel pixel NOUN fcis-29036 6 32 accuracy accuracy NOUN fcis-29036 6 33 ( ( PUNCT fcis-29036 6 34 mpa mpa PROPN fcis-29036 6 35 ) ) PUNCT fcis-29036 6 36 of of ADP fcis-29036 6 37 93.74 93.74 NUM fcis-29036 6 38 % % NOUN fcis-29036 6 39 . . PUNCT fcis-29036 7 1 the the DET fcis-29036 7 2 accuracy accuracy NOUN fcis-29036 7 3 for for ADP fcis-29036 7 4 impurity impurity NOUN fcis-29036 7 5 categories category NOUN fcis-29036 7 6 reaches reach VERB fcis-29036 7 7 95 95 NUM fcis-29036 7 8 % % NOUN fcis-29036 7 9 . . PUNCT fcis-29036 8 1 finally finally ADV fcis-29036 8 2 , , PUNCT fcis-29036 8 3 the the DET fcis-29036 8 4 model model NOUN fcis-29036 8 5 was be AUX fcis-29036 8 6 applied apply VERB fcis-29036 8 7 to to PART fcis-29036 8 8 test test VERB fcis-29036 8 9 the the DET fcis-29036 8 10 effectiveness effectiveness NOUN fcis-29036 8 11 in in ADP fcis-29036 8 12 identifying identify VERB fcis-29036 8 13 impurities impurity NOUN fcis-29036 8 14 ; ; PUNCT fcis-29036 8 15 the the DET fcis-29036 8 16 results result NOUN fcis-29036 8 17 indicate indicate VERB fcis-29036 8 18 that that SCONJ fcis-29036 8 19 the the DET fcis-29036 8 20 model model NOUN fcis-29036 8 21 can can AUX fcis-29036 8 22 effectively effectively ADV fcis-29036 8 23 identify identify VERB fcis-29036 8 24 various various ADJ fcis-29036 8 25 types type NOUN fcis-29036 8 26 of of ADP fcis-29036 8 27 impurities impurity NOUN fcis-29036 8 28 and and CCONJ fcis-29036 8 29 also also ADV fcis-29036 8 30 perform perform VERB fcis-29036 8 31 well well ADV fcis-29036 8 32 in in ADP fcis-29036 8 33 recognizing recognize VERB fcis-29036 8 34 unknown unknown ADJ fcis-29036 8 35 categories category NOUN fcis-29036 8 36 . . PUNCT fcis-29036 9 1 keywords keyword NOUN fcis-29036 9 2 : : PUNCT fcis-29036 9 3 multi multi ADJ fcis-29036 9 4 - - ADJ fcis-29036 9 5 modal modal ADJ fcis-29036 9 6 image image NOUN fcis-29036 9 7 fusion fusion NOUN fcis-29036 9 8 ; ; PUNCT fcis-29036 9 9 attention attention NOUN fcis-29036 9 10 unet unet NOUN fcis-29036 9 11 ; ; PUNCT fcis-29036 9 12 multispectral multispectral ADJ fcis-29036 9 13 images image NOUN fcis-29036 9 14 ; ; PUNCT fcis-29036 9 15 impurity impurity NOUN fcis-29036 9 16 detection detection NOUN fcis-29036 9 17 ; ; PUNCT fcis-29036 9 18 deep deep ADJ fcis-29036 9 19 learning learning NOUN fcis-29036 9 20 . . PUNCT fcis-29036 10 1 1 1 X fcis-29036 10 2 . . X fcis-29036 10 3 introduction introduction NOUN fcis-29036 10 4 food food NOUN fcis-29036 10 5 safety safety NOUN fcis-29036 10 6 and and CCONJ fcis-29036 10 7 quality quality NOUN fcis-29036 10 8 control control NOUN fcis-29036 10 9 are be AUX fcis-29036 10 10 critical critical ADJ fcis-29036 10 11 global global ADJ fcis-29036 10 12 concerns concern NOUN fcis-29036 10 13 . . PUNCT fcis-29036 11 1 as as SCONJ fcis-29036 11 2 food food NOUN fcis-29036 11 3 production production NOUN fcis-29036 11 4 processes process NOUN fcis-29036 11 5 and and CCONJ fcis-29036 11 6 supply supply NOUN fcis-29036 11 7 chains chain NOUN fcis-29036 11 8 become become VERB fcis-29036 11 9 increasingly increasingly ADV fcis-29036 11 10 complex complex ADJ fcis-29036 11 11 , , PUNCT fcis-29036 11 12 impurity impurity NOUN fcis-29036 11 13 detection detection NOUN fcis-29036 11 14 is be AUX fcis-29036 11 15 more more ADV fcis-29036 11 16 important important ADJ fcis-29036 11 17 than than ADP fcis-29036 11 18 ever ever ADV fcis-29036 11 19 . . PUNCT fcis-29036 12 1 traditional traditional ADJ fcis-29036 12 2 methods method NOUN fcis-29036 12 3 , , PUNCT fcis-29036 12 4 such such ADJ fcis-29036 12 5 as as ADP fcis-29036 12 6 physical physical ADJ fcis-29036 12 7 and and CCONJ fcis-29036 12 8 chemical chemical NOUN fcis-29036 12 9 laboratory laboratory NOUN fcis-29036 12 10 analyses analysis NOUN fcis-29036 12 11 , , PUNCT fcis-29036 12 12 are be AUX fcis-29036 12 13 accurate accurate ADJ fcis-29036 12 14 but but CCONJ fcis-29036 12 15 time time NOUN fcis-29036 12 16 - - PUNCT fcis-29036 12 17 consuming consume VERB fcis-29036 12 18 and and CCONJ fcis-29036 12 19 destructive destructive ADJ fcis-29036 12 20 , , PUNCT fcis-29036 12 21 making make VERB fcis-29036 12 22 them they PRON fcis-29036 12 23 unsuitable unsuitable ADJ fcis-29036 12 24 for for ADP fcis-29036 12 25 rapid rapid ADJ fcis-29036 12 26 and and CCONJ fcis-29036 12 27 convenient convenient ADJ fcis-29036 12 28 applications application NOUN fcis-29036 12 29 [ [ X fcis-29036 12 30 1 1 NUM fcis-29036 12 31 ] ] PUNCT fcis-29036 12 32 . . PUNCT fcis-29036 13 1 as as ADP fcis-29036 13 2 one one NUM fcis-29036 13 3 of of ADP fcis-29036 13 4 the the DET fcis-29036 13 5 most most ADV fcis-29036 13 6 commonly commonly ADV fcis-29036 13 7 used use VERB fcis-29036 13 8 sweeteners sweetener NOUN fcis-29036 13 9 [ [ X fcis-29036 13 10 2 2 NUM fcis-29036 13 11 ] ] PUNCT fcis-29036 13 12 , , PUNCT fcis-29036 13 13 sugar sugar NOUN fcis-29036 13 14 is be AUX fcis-29036 13 15 widely widely ADV fcis-29036 13 16 applied apply VERB fcis-29036 13 17 in in ADP fcis-29036 13 18 the the DET fcis-29036 13 19 food food NOUN fcis-29036 13 20 and and CCONJ fcis-29036 13 21 beverage beverage NOUN fcis-29036 13 22 industries industry NOUN fcis-29036 13 23 . . PUNCT fcis-29036 14 1 the the DET fcis-29036 14 2 sugar sugar NOUN fcis-29036 14 3 available available ADJ fcis-29036 14 4 on on ADP fcis-29036 14 5 the the DET fcis-29036 14 6 market market NOUN fcis-29036 14 7 today today NOUN fcis-29036 14 8 is be AUX fcis-29036 14 9 primarily primarily ADV fcis-29036 14 10 derived derive VERB fcis-29036 14 11 from from ADP fcis-29036 14 12 sugarcane sugarcane NOUN fcis-29036 14 13 or or CCONJ fcis-29036 14 14 sugar sugar NOUN fcis-29036 14 15 beet beet NOUN fcis-29036 14 16 . . PUNCT fcis-29036 15 1 during during ADP fcis-29036 15 2 production production NOUN fcis-29036 15 3 and and CCONJ fcis-29036 15 4 storage storage NOUN fcis-29036 15 5 , , PUNCT fcis-29036 15 6 sugar sugar NOUN fcis-29036 15 7 is be AUX fcis-29036 15 8 prone prone ADJ fcis-29036 15 9 to to ADP fcis-29036 15 10 contamination contamination NOUN fcis-29036 15 11 by by ADP fcis-29036 15 12 impurities impurity NOUN fcis-29036 15 13 . . PUNCT fcis-29036 16 1 traditional traditional ADJ fcis-29036 16 2 impurity impurity NOUN fcis-29036 16 3 removal removal NOUN fcis-29036 16 4 methods method NOUN fcis-29036 16 5 rely rely VERB fcis-29036 16 6 on on ADP fcis-29036 16 7 the the DET fcis-29036 16 8 sugar sugar NOUN fcis-29036 16 9 ’s ’s PART fcis-29036 16 10 water water NOUN fcis-29036 16 11 solubility solubility NOUN fcis-29036 16 12 , , PUNCT fcis-29036 16 13 requiring require VERB fcis-29036 16 14 repeated repeat VERB fcis-29036 16 15 processing processing NOUN fcis-29036 16 16 with with ADP fcis-29036 16 17 limited limited ADJ fcis-29036 16 18 effectiveness effectiveness NOUN fcis-29036 16 19 . . PUNCT fcis-29036 17 1 remaining remain VERB fcis-29036 17 2 impurities impurity NOUN fcis-29036 17 3 often often ADV fcis-29036 17 4 require require VERB fcis-29036 17 5 manual manual ADJ fcis-29036 17 6 sorting sorting NOUN fcis-29036 17 7 , , PUNCT fcis-29036 17 8 resulting result VERB fcis-29036 17 9 in in ADP fcis-29036 17 10 high high ADJ fcis-29036 17 11 detection detection NOUN fcis-29036 17 12 costs cost NOUN fcis-29036 17 13 and and CCONJ fcis-29036 17 14 inefficiency inefficiency NOUN fcis-29036 17 15 . . PUNCT fcis-29036 18 1 research research NOUN fcis-29036 18 2 on on ADP fcis-29036 18 3 sugar sugar NOUN fcis-29036 18 4 has have AUX fcis-29036 18 5 mainly mainly ADV fcis-29036 18 6 focused focus VERB fcis-29036 18 7 on on ADP fcis-29036 18 8 qualitative qualitative ADJ fcis-29036 18 9 analysis analysis NOUN fcis-29036 18 10 , , PUNCT fcis-29036 18 11 sugar sugar NOUN fcis-29036 18 12 content content NOUN fcis-29036 18 13 , , PUNCT fcis-29036 18 14 and and CCONJ fcis-29036 18 15 particle particle NOUN fcis-29036 18 16 size size NOUN fcis-29036 18 17 measurement measurement NOUN fcis-29036 19 1 [ [ X fcis-29036 19 2 3,4 3,4 NUM fcis-29036 19 3 ] ] PUNCT fcis-29036 19 4 , , PUNCT fcis-29036 19 5 with with ADP fcis-29036 19 6 fewer few ADJ fcis-29036 19 7 studies study NOUN fcis-29036 19 8 addressing address VERB fcis-29036 19 9 impurity impurity NOUN fcis-29036 19 10 detection detection NOUN fcis-29036 19 11 . . PUNCT fcis-29036 20 1 spectral spectral ADJ fcis-29036 20 2 vision vision NOUN fcis-29036 20 3 , , PUNCT fcis-29036 20 4 which which PRON fcis-29036 20 5 combines combine VERB fcis-29036 20 6 imaging imaging NOUN fcis-29036 20 7 and and CCONJ fcis-29036 20 8 spectroscopy spectroscopy NOUN fcis-29036 20 9 , , PUNCT fcis-29036 20 10 enables enable VERB fcis-29036 20 11 simultaneous simultaneous ADJ fcis-29036 20 12 acquisition acquisition NOUN fcis-29036 20 13 of of ADP fcis-29036 20 14 spatial spatial ADJ fcis-29036 20 15 and and CCONJ fcis-29036 20 16 spectral spectral ADJ fcis-29036 20 17 information information NOUN fcis-29036 20 18 from from ADP fcis-29036 20 19 samples sample NOUN fcis-29036 20 20 . . PUNCT fcis-29036 21 1 depending depend VERB fcis-29036 21 2 on on ADP fcis-29036 21 3 resolution resolution NOUN fcis-29036 21 4 , , PUNCT fcis-29036 21 5 it it PRON fcis-29036 21 6 can can AUX fcis-29036 21 7 be be AUX fcis-29036 21 8 categorized categorize VERB fcis-29036 21 9 into into ADP fcis-29036 21 10 hyperspectral hyperspectral ADJ fcis-29036 21 11 and and CCONJ fcis-29036 21 12 multispectral multispectral ADJ fcis-29036 21 13 imaging imaging NOUN fcis-29036 21 14 . . PUNCT fcis-29036 22 1 this this DET fcis-29036 22 2 technology technology NOUN fcis-29036 22 3 not not PART fcis-29036 22 4 only only ADV fcis-29036 22 5 provides provide VERB fcis-29036 22 6 detailed detailed ADJ fcis-29036 22 7 spatial spatial ADJ fcis-29036 22 8 information information NOUN fcis-29036 22 9 but but CCONJ fcis-29036 22 10 also also ADV fcis-29036 22 11 reveals reveal VERB fcis-29036 22 12 the the DET fcis-29036 22 13 chemical chemical NOUN fcis-29036 22 14 composition composition NOUN fcis-29036 22 15 of of ADP fcis-29036 22 16 materials material NOUN fcis-29036 22 17 , , PUNCT fcis-29036 22 18 making make VERB fcis-29036 22 19 it it PRON fcis-29036 22 20 highly highly ADV fcis-29036 22 21 suitable suitable ADJ fcis-29036 22 22 for for ADP fcis-29036 22 23 food food NOUN fcis-29036 22 24 detection detection NOUN fcis-29036 22 25 . . PUNCT fcis-29036 23 1 for for ADP fcis-29036 23 2 example example NOUN fcis-29036 23 3 , , PUNCT fcis-29036 23 4 multispectral multispectral ADJ fcis-29036 23 5 imaging imaging NOUN fcis-29036 23 6 has have AUX fcis-29036 23 7 been be AUX fcis-29036 23 8 used use VERB fcis-29036 23 9 to to PART fcis-29036 23 10 detect detect VERB fcis-29036 23 11 mango mango NOUN fcis-29036 23 12 chilling chilling ADJ fcis-29036 23 13 damage damage NOUN fcis-29036 24 1 [ [ X fcis-29036 24 2 5 5 NUM fcis-29036 24 3 ] ] PUNCT fcis-29036 24 4 and and CCONJ fcis-29036 24 5 non non ADJ fcis-29036 24 6 - - ADJ fcis-29036 24 7 smoking smoking ADJ fcis-29036 24 8 materials material NOUN fcis-29036 24 9 in in ADP fcis-29036 24 10 cigarettes cigarette NOUN fcis-29036 24 11 [ [ X fcis-29036 24 12 6 6 NUM fcis-29036 24 13 ] ] PUNCT fcis-29036 24 14 , , PUNCT fcis-29036 24 15 achieving achieve VERB fcis-29036 24 16 high high ADJ fcis-29036 24 17 detection detection NOUN fcis-29036 24 18 accuracy accuracy NOUN fcis-29036 24 19 in in ADP fcis-29036 24 20 complex complex ADJ fcis-29036 24 21 scenarios scenario NOUN fcis-29036 24 22 . . PUNCT fcis-29036 25 1 based base VERB fcis-29036 25 2 on on ADP fcis-29036 25 3 the the DET fcis-29036 25 4 background background NOUN fcis-29036 25 5 , , PUNCT fcis-29036 25 6 this this DET fcis-29036 25 7 study study NOUN fcis-29036 25 8 proposes propose VERB fcis-29036 25 9 an an DET fcis-29036 25 10 automated automate VERB fcis-29036 25 11 sugar sugar NOUN fcis-29036 25 12 impurity impurity NOUN fcis-29036 25 13 detection detection NOUN fcis-29036 25 14 system system NOUN fcis-29036 25 15 using use VERB fcis-29036 25 16 multimodal multimodal NOUN fcis-29036 25 17 image image NOUN fcis-29036 25 18 fusion fusion NOUN fcis-29036 25 19 technology technology NOUN fcis-29036 25 20 . . PUNCT fcis-29036 26 1 the the DET fcis-29036 26 2 system system NOUN fcis-29036 26 3 employs employ VERB fcis-29036 26 4 a a DET fcis-29036 26 5 multispectral multispectral NOUN fcis-29036 26 6 and and CCONJ fcis-29036 26 7 an an DET fcis-29036 26 8 rgb rgb PROPN fcis-29036 26 9 cameras camera NOUN fcis-29036 26 10 for for ADP fcis-29036 26 11 image image NOUN fcis-29036 26 12 capture capture NOUN fcis-29036 26 13 , , PUNCT fcis-29036 26 14 along along ADP fcis-29036 26 15 with with ADP fcis-29036 26 16 a a DET fcis-29036 26 17 deep deep ADJ fcis-29036 26 18 learning learning NOUN fcis-29036 26 19 model model NOUN fcis-29036 26 20 ( ( PUNCT fcis-29036 26 21 attention attention NOUN fcis-29036 26 22 u u NOUN fcis-29036 26 23 - - NOUN fcis-29036 26 24 net net NOUN fcis-29036 26 25 ) ) PUNCT fcis-29036 26 26 for for ADP fcis-29036 26 27 automated automate VERB fcis-29036 26 28 impurity impurity NOUN fcis-29036 26 29 detection detection NOUN fcis-29036 26 30 . . PUNCT fcis-29036 27 1 multispectral multispectral ADJ fcis-29036 27 2 imaging imaging NOUN fcis-29036 27 3 enables enable VERB fcis-29036 27 4 precise precise ADJ fcis-29036 27 5 composition composition NOUN fcis-29036 27 6 analysis analysis NOUN fcis-29036 27 7 , , PUNCT fcis-29036 27 8 while while SCONJ fcis-29036 27 9 rgb rgb PROPN fcis-29036 27 10 imaging imaging NOUN fcis-29036 27 11 captures capture VERB fcis-29036 27 12 object object VERB fcis-29036 27 13 appearances appearance NOUN fcis-29036 27 14 . . PUNCT fcis-29036 28 1 compared compare VERB fcis-29036 28 2 to to ADP fcis-29036 28 3 traditional traditional ADJ fcis-29036 28 4 methods method NOUN fcis-29036 28 5 , , PUNCT fcis-29036 28 6 this this DET fcis-29036 28 7 system system NOUN fcis-29036 28 8 offers offer VERB fcis-29036 28 9 significant significant ADJ fcis-29036 28 10 advantages advantage NOUN fcis-29036 28 11 in in ADP fcis-29036 28 12 non non ADJ fcis-29036 28 13 - - ADJ fcis-29036 28 14 destructive destructive ADJ fcis-29036 28 15 testing testing NOUN fcis-29036 28 16 and and CCONJ fcis-29036 28 17 realtime realtime VERB fcis-29036 28 18 analysis analysis NOUN fcis-29036 28 19 , , PUNCT fcis-29036 28 20 improving improve VERB fcis-29036 28 21 detection detection NOUN fcis-29036 28 22 efficiency efficiency NOUN fcis-29036 28 23 and and CCONJ fcis-29036 28 24 accuracy accuracy NOUN fcis-29036 28 25 . . PUNCT fcis-29036 29 1 2 2 X fcis-29036 29 2 . . X fcis-29036 29 3 overall overall ADJ fcis-29036 29 4 framework framework NOUN fcis-29036 29 5 design design NOUN fcis-29036 29 6 of of ADP fcis-29036 29 7 the the DET fcis-29036 29 8 system system NOUN fcis-29036 29 9 2.1 2.1 NUM fcis-29036 29 10 . . PUNCT fcis-29036 30 1 functional functional ADJ fcis-29036 30 2 design design NOUN fcis-29036 30 3 of of ADP fcis-29036 30 4 the the DET fcis-29036 30 5 system system NOUN fcis-29036 30 6 this this DET fcis-29036 30 7 study study NOUN fcis-29036 30 8 develops develop VERB fcis-29036 30 9 an an DET fcis-29036 30 10 automated automate VERB fcis-29036 30 11 sugar sugar NOUN fcis-29036 30 12 impurity impurity NOUN fcis-29036 30 13 detection detection NOUN fcis-29036 30 14 system system NOUN fcis-29036 30 15 based base VERB fcis-29036 30 16 on on ADP fcis-29036 30 17 spectral spectral ADJ fcis-29036 30 18 imaging imaging NOUN fcis-29036 30 19 and and CCONJ fcis-29036 30 20 deep deep ADJ fcis-29036 30 21 learning learning NOUN fcis-29036 30 22 , , PUNCT fcis-29036 30 23 achieving achieve VERB fcis-29036 30 24 efficient efficient ADJ fcis-29036 30 25 , , PUNCT fcis-29036 30 26 accurate accurate ADJ fcis-29036 30 27 , , PUNCT fcis-29036 30 28 and and CCONJ fcis-29036 30 29 real real ADJ fcis-29036 30 30 - - PUNCT fcis-29036 30 31 time time NOUN fcis-29036 30 32 detection detection NOUN fcis-29036 30 33 to to PART fcis-29036 30 34 meet meet VERB fcis-29036 30 35 the the DET fcis-29036 30 36 quality quality NOUN fcis-29036 30 37 control control NOUN fcis-29036 30 38 demands demand NOUN fcis-29036 30 39 in in ADP fcis-29036 30 40 complex complex ADJ fcis-29036 30 41 production production NOUN fcis-29036 30 42 environments environment NOUN fcis-29036 30 43 . . PUNCT fcis-29036 31 1 the the DET fcis-29036 31 2 system system NOUN fcis-29036 31 3 integrates integrate VERB fcis-29036 31 4 multispectral multispectral ADJ fcis-29036 31 5 image image NOUN fcis-29036 31 6 processing processing NOUN fcis-29036 31 7 techniques technique NOUN fcis-29036 31 8 and and CCONJ fcis-29036 31 9 deep deep ADJ fcis-29036 31 10 learning learning NOUN fcis-29036 31 11 models model NOUN fcis-29036 31 12 , , PUNCT fcis-29036 31 13 enabling enable VERB fcis-29036 31 14 rapid rapid ADJ fcis-29036 31 15 impurity impurity NOUN fcis-29036 31 16 identification identification NOUN fcis-29036 31 17 under under ADP fcis-29036 31 18 complex complex ADJ fcis-29036 31 19 backgrounds background NOUN fcis-29036 31 20 while while SCONJ fcis-29036 31 21 ensuring ensure VERB fcis-29036 31 22 the the DET fcis-29036 31 23 stability stability NOUN fcis-29036 31 24 and and CCONJ fcis-29036 31 25 accuracy accuracy NOUN fcis-29036 31 26 of of ADP fcis-29036 31 27 detection detection NOUN fcis-29036 31 28 results result NOUN fcis-29036 31 29 . . PUNCT fcis-29036 32 1 the the DET fcis-29036 32 2 system system NOUN fcis-29036 32 3 is be AUX fcis-29036 32 4 user user NOUN fcis-29036 32 5 - - PUNCT fcis-29036 32 6 friendly friendly ADJ fcis-29036 32 7 , , PUNCT fcis-29036 32 8 requiring require VERB fcis-29036 32 9 no no DET fcis-29036 32 10 specialized specialized ADJ fcis-29036 32 11 knowledge knowledge NOUN fcis-29036 32 12 for for ADP fcis-29036 32 13 operation operation NOUN fcis-29036 32 14 , , PUNCT fcis-29036 32 15 and and CCONJ fcis-29036 32 16 offers offer VERB fcis-29036 32 17 flexibility flexibility NOUN fcis-29036 32 18 and and CCONJ fcis-29036 32 19 scalability scalability NOUN fcis-29036 32 20 for for ADP fcis-29036 32 21 future future ADJ fcis-29036 32 22 upgrades upgrade NOUN fcis-29036 32 23 in in ADP fcis-29036 32 24 line line NOUN fcis-29036 32 25 with with ADP fcis-29036 32 26 technological technological ADJ fcis-29036 32 27 advancements advancement NOUN fcis-29036 32 28 and and CCONJ fcis-29036 32 29 market market NOUN fcis-29036 32 30 demands demand NOUN fcis-29036 32 31 . . PUNCT fcis-29036 33 1 the the DET fcis-29036 33 2 system system NOUN fcis-29036 33 3 comprises comprise VERB fcis-29036 33 4 the the DET fcis-29036 33 5 following follow VERB fcis-29036 33 6 key key ADJ fcis-29036 33 7 modules module NOUN fcis-29036 33 8 , , PUNCT fcis-29036 33 9 with with SCONJ fcis-29036 33 10 the the DET fcis-29036 33 11 system system NOUN fcis-29036 33 12 architecture architecture NOUN fcis-29036 33 13 illustrated illustrate VERB fcis-29036 33 14 in in ADP fcis-29036 33 15 figure figure NOUN fcis-29036 33 16 1 1 NUM fcis-29036 33 17 : : PUNCT fcis-29036 33 18 ( ( PUNCT fcis-29036 33 19 1 1 X fcis-29036 33 20 ) ) PUNCT fcis-29036 33 21 imaging imaging NOUN fcis-29036 33 22 system system NOUN fcis-29036 33 23 : : PUNCT fcis-29036 33 24 before before ADP fcis-29036 33 25 operation operation NOUN fcis-29036 33 26 , , PUNCT fcis-29036 33 27 the the DET fcis-29036 33 28 camera camera NOUN fcis-29036 33 29 's 's PART fcis-29036 33 30 exposure exposure NOUN fcis-29036 33 31 time time NOUN fcis-29036 33 32 and and CCONJ fcis-29036 33 33 focal focal ADJ fcis-29036 33 34 length length NOUN fcis-29036 33 35 need need VERB fcis-29036 33 36 to to PART fcis-29036 33 37 be be AUX fcis-29036 33 38 adjusted adjust VERB fcis-29036 33 39 and and CCONJ fcis-29036 33 40 set set VERB fcis-29036 33 41 appropriately appropriately ADV fcis-29036 33 42 . . PUNCT fcis-29036 34 1 the the DET fcis-29036 34 2 image image NOUN fcis-29036 34 3 capture capture NOUN fcis-29036 34 4 process process NOUN fcis-29036 34 5 can can AUX fcis-29036 34 6 be be AUX fcis-29036 34 7 controlled control VERB fcis-29036 34 8 via via ADP fcis-29036 34 9 software software NOUN fcis-29036 34 10 , , PUNCT fcis-29036 34 11 allowing allow VERB fcis-29036 34 12 the the DET fcis-29036 34 13 system system NOUN fcis-29036 34 14 to to PART fcis-29036 34 15 save save VERB fcis-29036 34 16 the the DET fcis-29036 34 17 acquired acquire VERB fcis-29036 34 18 image image NOUN fcis-29036 34 19 data datum NOUN fcis-29036 34 20 . . PUNCT fcis-29036 35 1 during during ADP fcis-29036 35 2 system system NOUN fcis-29036 35 3 6 6 NUM fcis-29036 35 4 operation operation NOUN fcis-29036 35 5 , , PUNCT fcis-29036 35 6 real real ADJ fcis-29036 35 7 - - PUNCT fcis-29036 35 8 time time NOUN fcis-29036 35 9 image image NOUN fcis-29036 35 10 acquisition acquisition NOUN fcis-29036 35 11 from from ADP fcis-29036 35 12 the the DET fcis-29036 35 13 conveyor conveyor NOUN fcis-29036 35 14 belt belt NOUN fcis-29036 35 15 is be AUX fcis-29036 35 16 required require VERB fcis-29036 35 17 . . PUNCT fcis-29036 36 1 the the DET fcis-29036 36 2 software software NOUN fcis-29036 36 3 is be AUX fcis-29036 36 4 capable capable ADJ fcis-29036 36 5 of of ADP fcis-29036 36 6 displaying display VERB fcis-29036 36 7 both both CCONJ fcis-29036 36 8 the the DET fcis-29036 36 9 captured capture VERB fcis-29036 36 10 images image NOUN fcis-29036 36 11 and and CCONJ fcis-29036 36 12 the the DET fcis-29036 36 13 results result NOUN fcis-29036 36 14 of of ADP fcis-29036 36 15 model model NOUN fcis-29036 36 16 analysis analysis NOUN fcis-29036 36 17 in in ADP fcis-29036 36 18 real real ADJ fcis-29036 36 19 time time NOUN fcis-29036 36 20 , , PUNCT fcis-29036 36 21 providing provide VERB fcis-29036 36 22 continuous continuous ADJ fcis-29036 36 23 monitoring monitoring NOUN fcis-29036 36 24 and and CCONJ fcis-29036 36 25 feedback feedback NOUN fcis-29036 36 26 throughout throughout ADP fcis-29036 36 27 the the DET fcis-29036 36 28 detection detection NOUN fcis-29036 36 29 process process NOUN fcis-29036 36 30 . . PUNCT fcis-29036 37 1 ( ( PUNCT fcis-29036 37 2 2 2 X fcis-29036 37 3 ) ) PUNCT fcis-29036 37 4 computer computer NOUN fcis-29036 37 5 : : PUNCT fcis-29036 37 6 the the DET fcis-29036 37 7 computer computer NOUN fcis-29036 37 8 serves serve VERB fcis-29036 37 9 as as ADP fcis-29036 37 10 the the DET fcis-29036 37 11 master master PROPN fcis-29036 37 12 control control PROPN fcis-29036 37 13 unit unit NOUN fcis-29036 37 14 , , PUNCT fcis-29036 37 15 allowing allow VERB fcis-29036 37 16 software software NOUN fcis-29036 37 17 - - PUNCT fcis-29036 37 18 based base VERB fcis-29036 37 19 control control NOUN fcis-29036 37 20 of of ADP fcis-29036 37 21 the the DET fcis-29036 37 22 conveyor conveyor NOUN fcis-29036 37 23 belt belt NOUN fcis-29036 37 24 system system NOUN fcis-29036 37 25 , , PUNCT fcis-29036 37 26 image image NOUN fcis-29036 37 27 acquisition acquisition NOUN fcis-29036 37 28 , , PUNCT fcis-29036 37 29 and and CCONJ fcis-29036 37 30 display display NOUN fcis-29036 37 31 of of ADP fcis-29036 37 32 the the DET fcis-29036 37 33 model model NOUN fcis-29036 37 34 processing processing NOUN fcis-29036 37 35 results result NOUN fcis-29036 37 36 . . PUNCT fcis-29036 38 1 it it PRON fcis-29036 38 2 manages manage VERB fcis-29036 38 3 the the DET fcis-29036 38 4 entire entire ADJ fcis-29036 38 5 detection detection NOUN fcis-29036 38 6 process process NOUN fcis-29036 38 7 , , PUNCT fcis-29036 38 8 facilitating facilitate VERB fcis-29036 38 9 seamless seamless ADJ fcis-29036 38 10 coordination coordination NOUN fcis-29036 38 11 between between ADP fcis-29036 38 12 the the DET fcis-29036 38 13 conveyor conveyor NOUN fcis-29036 38 14 belt belt NOUN fcis-29036 38 15 , , PUNCT fcis-29036 38 16 image image NOUN fcis-29036 38 17 capture capture NOUN fcis-29036 38 18 , , PUNCT fcis-29036 38 19 and and CCONJ fcis-29036 38 20 real real ADJ fcis-29036 38 21 - - PUNCT fcis-29036 38 22 time time NOUN fcis-29036 38 23 presentation presentation NOUN fcis-29036 38 24 of of ADP fcis-29036 38 25 analysis analysis NOUN fcis-29036 38 26 outcomes outcome NOUN fcis-29036 38 27 . . PUNCT fcis-29036 39 1 ( ( PUNCT fcis-29036 39 2 3 3 X fcis-29036 39 3 ) ) PUNCT fcis-29036 39 4 plc plc PROPN fcis-29036 39 5 control control NOUN fcis-29036 39 6 system system NOUN fcis-29036 39 7 : : PUNCT fcis-29036 39 8 the the DET fcis-29036 39 9 programmable programmable ADJ fcis-29036 39 10 logic logic NOUN fcis-29036 39 11 controller controller NOUN fcis-29036 39 12 ( ( PUNCT fcis-29036 39 13 plc plc NOUN fcis-29036 39 14 ) ) PUNCT fcis-29036 39 15 serves serve VERB fcis-29036 39 16 as as ADP fcis-29036 39 17 the the DET fcis-29036 39 18 slave slave NOUN fcis-29036 39 19 unit unit NOUN fcis-29036 39 20 , , PUNCT fcis-29036 39 21 communicating communicate VERB fcis-29036 39 22 with with ADP fcis-29036 39 23 the the DET fcis-29036 39 24 master master PROPN fcis-29036 39 25 control control PROPN fcis-29036 39 26 unit unit NOUN fcis-29036 39 27 ( ( PUNCT fcis-29036 39 28 such such ADJ fcis-29036 39 29 as as ADP fcis-29036 39 30 a a DET fcis-29036 39 31 computer computer NOUN fcis-29036 39 32 ) ) PUNCT fcis-29036 39 33 . . PUNCT fcis-29036 40 1 it it PRON fcis-29036 40 2 receives receive VERB fcis-29036 40 3 command command NOUN fcis-29036 40 4 signals signal NOUN fcis-29036 40 5 from from ADP fcis-29036 40 6 the the DET fcis-29036 40 7 master master NOUN fcis-29036 40 8 unit unit NOUN fcis-29036 40 9 and and CCONJ fcis-29036 40 10 executes execute VERB fcis-29036 40 11 the the DET fcis-29036 40 12 corresponding correspond VERB fcis-29036 40 13 control control NOUN fcis-29036 40 14 functions function NOUN fcis-29036 40 15 based base VERB fcis-29036 40 16 on on ADP fcis-29036 40 17 these these DET fcis-29036 40 18 instructions instruction NOUN fcis-29036 40 19 , , PUNCT fcis-29036 40 20 transmitting transmit VERB fcis-29036 40 21 control control NOUN fcis-29036 40 22 signals signal NOUN fcis-29036 40 23 to to ADP fcis-29036 40 24 the the DET fcis-29036 40 25 subordinate subordinate ADJ fcis-29036 40 26 devices device NOUN fcis-29036 40 27 . . PUNCT fcis-29036 41 1 ( ( PUNCT fcis-29036 41 2 4 4 X fcis-29036 41 3 ) ) PUNCT fcis-29036 41 4 conveyor conveyor NOUN fcis-29036 41 5 belt belt NOUN fcis-29036 41 6 system system NOUN fcis-29036 41 7 : : PUNCT fcis-29036 41 8 to to PART fcis-29036 41 9 achieve achieve VERB fcis-29036 41 10 automated automate VERB fcis-29036 41 11 and and CCONJ fcis-29036 41 12 real real ADJ fcis-29036 41 13 - - PUNCT fcis-29036 41 14 time time NOUN fcis-29036 41 15 impurity impurity NOUN fcis-29036 41 16 detection detection NOUN fcis-29036 41 17 , , PUNCT fcis-29036 41 18 the the DET fcis-29036 41 19 system system NOUN fcis-29036 41 20 must must AUX fcis-29036 41 21 ensure ensure VERB fcis-29036 41 22 that that SCONJ fcis-29036 41 23 samples sample NOUN fcis-29036 41 24 are be AUX fcis-29036 41 25 detected detect VERB fcis-29036 41 26 while while SCONJ fcis-29036 41 27 dynamically dynamically ADV fcis-29036 41 28 moving move VERB fcis-29036 41 29 along along ADP fcis-29036 41 30 the the DET fcis-29036 41 31 conveyor conveyor NOUN fcis-29036 41 32 belt belt NOUN fcis-29036 41 33 . . PUNCT fcis-29036 42 1 accurate accurate ADJ fcis-29036 42 2 control control NOUN fcis-29036 42 3 of of ADP fcis-29036 42 4 the the DET fcis-29036 42 5 conveyor conveyor NOUN fcis-29036 42 6 belt belt NOUN fcis-29036 42 7 's 's PART fcis-29036 42 8 speed speed NOUN fcis-29036 42 9 and and CCONJ fcis-29036 42 10 direction direction NOUN fcis-29036 42 11 is be AUX fcis-29036 42 12 crucial crucial ADJ fcis-29036 42 13 , , PUNCT fcis-29036 42 14 as as SCONJ fcis-29036 42 15 it it PRON fcis-29036 42 16 not not PART fcis-29036 42 17 only only ADV fcis-29036 42 18 affects affect VERB fcis-29036 42 19 data datum NOUN fcis-29036 42 20 acquisition acquisition NOUN fcis-29036 42 21 efficiency efficiency NOUN fcis-29036 42 22 but but CCONJ fcis-29036 42 23 also also ADV fcis-29036 42 24 plays play VERB fcis-29036 42 25 a a DET fcis-29036 42 26 key key ADJ fcis-29036 42 27 role role NOUN fcis-29036 42 28 in in ADP fcis-29036 42 29 the the DET fcis-29036 42 30 subsequent subsequent ADJ fcis-29036 42 31 impurity impurity NOUN fcis-29036 42 32 removal removal NOUN fcis-29036 42 33 process process NOUN fcis-29036 42 34 . . PUNCT fcis-29036 43 1 the the DET fcis-29036 43 2 conveyor conveyor NOUN fcis-29036 43 3 belt belt NOUN fcis-29036 43 4 parameters parameter NOUN fcis-29036 43 5 must must AUX fcis-29036 43 6 be be AUX fcis-29036 43 7 carefully carefully ADV fcis-29036 43 8 set set VERB fcis-29036 43 9 to to PART fcis-29036 43 10 meet meet VERB fcis-29036 43 11 the the DET fcis-29036 43 12 demands demand NOUN fcis-29036 43 13 of of ADP fcis-29036 43 14 high high ADJ fcis-29036 43 15 - - PUNCT fcis-29036 43 16 speed speed NOUN fcis-29036 43 17 , , PUNCT fcis-29036 43 18 high high ADJ fcis-29036 43 19 - - PUNCT fcis-29036 43 20 quality quality NOUN fcis-29036 43 21 data datum NOUN fcis-29036 43 22 collection collection NOUN fcis-29036 43 23 while while SCONJ fcis-29036 43 24 providing provide VERB fcis-29036 43 25 sufficient sufficient ADJ fcis-29036 43 26 time time NOUN fcis-29036 43 27 and and CCONJ fcis-29036 43 28 space space NOUN fcis-29036 43 29 for for ADP fcis-29036 43 30 impurity impurity NOUN fcis-29036 43 31 selection selection NOUN fcis-29036 43 32 , , PUNCT fcis-29036 43 33 thereby thereby ADV fcis-29036 43 34 improving improve VERB fcis-29036 43 35 the the DET fcis-29036 43 36 overall overall ADJ fcis-29036 43 37 efficiency efficiency NOUN fcis-29036 43 38 and and CCONJ fcis-29036 43 39 accuracy accuracy NOUN fcis-29036 43 40 of of ADP fcis-29036 43 41 the the DET fcis-29036 43 42 production production NOUN fcis-29036 43 43 line line NOUN fcis-29036 43 44 . . PUNCT fcis-29036 44 1 rgb rgb PROPN fcis-29036 44 2 , , PUNCT fcis-29036 44 3     SPACE fcis-29036 44 4 multispectral multispectral ADJ fcis-29036 44 5     SPACE fcis-29036 44 6 imaging imaging NOUN fcis-29036 44 7     SPACE fcis-29036 44 8 system system NOUN fcis-29036 44 9 computer computer NOUN fcis-29036 44 10     SPACE fcis-29036 44 11 pcl pcl PROPN fcis-29036 44 12     SPACE fcis-29036 44 13 control control NOUN fcis-29036 44 14     SPACE fcis-29036 44 15 system system NOUN fcis-29036 44 16 conveyor conveyor NOUN fcis-29036 44 17     SPACE fcis-29036 44 18 belt belt NOUN fcis-29036 44 19     SPACE fcis-29036 44 20 system system NOUN fcis-29036 44 21 other other ADJ fcis-29036 44 22     SPACE fcis-29036 44 23 modules module NOUN fcis-29036 44 24 fig fig NOUN fcis-29036 44 25 1 1 NUM fcis-29036 44 26 . . PUNCT fcis-29036 44 27 system system NOUN fcis-29036 44 28 architecture architecture NOUN fcis-29036 44 29 diagram diagram NOUN fcis-29036 44 30 2.2 2.2 NUM fcis-29036 44 31 . . PUNCT fcis-29036 45 1 system system NOUN fcis-29036 45 2 hardware hardware NOUN fcis-29036 45 3 design design NOUN fcis-29036 45 4 in in ADP fcis-29036 45 5 this this DET fcis-29036 45 6 study study NOUN fcis-29036 45 7 , , PUNCT fcis-29036 45 8 spectral spectral ADJ fcis-29036 45 9 image image NOUN fcis-29036 45 10 data datum NOUN fcis-29036 45 11 of of ADP fcis-29036 45 12 sugar sugar NOUN fcis-29036 45 13 and and CCONJ fcis-29036 45 14 various various ADJ fcis-29036 45 15 impurities impurity NOUN fcis-29036 45 16 were be AUX fcis-29036 45 17 collected collect VERB fcis-29036 45 18 using use VERB fcis-29036 45 19 the the DET fcis-29036 45 20 spectral spectral ADJ fcis-29036 45 21 system system NOUN fcis-29036 45 22 shown show VERB fcis-29036 45 23 in in ADP fcis-29036 45 24 figure figure NOUN fcis-29036 45 25 2 2 NUM fcis-29036 45 26 . . PUNCT fcis-29036 46 1 the the DET fcis-29036 46 2 system system NOUN fcis-29036 46 3 consists consist VERB fcis-29036 46 4 of of ADP fcis-29036 46 5 four four NUM fcis-29036 46 6 multispectral multispectral ADJ fcis-29036 46 7 cameras camera NOUN fcis-29036 46 8 and and CCONJ fcis-29036 46 9 an an DET fcis-29036 46 10 rgb rgb PROPN fcis-29036 46 11 camera camera NOUN fcis-29036 46 12 . . PUNCT fcis-29036 47 1 the the DET fcis-29036 47 2 multispectral multispectral ADJ fcis-29036 47 3 cameras camera NOUN fcis-29036 47 4 are be AUX fcis-29036 47 5 designed design VERB fcis-29036 47 6 with with ADP fcis-29036 47 7 ten ten NUM fcis-29036 47 8 bands band NOUN fcis-29036 47 9 , , PUNCT fcis-29036 47 10 covering cover VERB fcis-29036 47 11 wavelengths wavelength NOUN fcis-29036 47 12 from from ADP fcis-29036 47 13 713 713 NUM fcis-29036 47 14 nm nm NOUN fcis-29036 47 15 to to ADP fcis-29036 47 16 920 920 NUM fcis-29036 47 17 nm nm NOUN fcis-29036 47 18 , , PUNCT fcis-29036 47 19 with with ADP fcis-29036 47 20 a a DET fcis-29036 47 21 wavelength wavelength NOUN fcis-29036 47 22 resolution resolution NOUN fcis-29036 47 23 of of ADP fcis-29036 47 24 23 23 NUM fcis-29036 47 25 nm nm NOUN fcis-29036 47 26 . . PUNCT fcis-29036 48 1 their their PRON fcis-29036 48 2 compact compact ADJ fcis-29036 48 3 size size NOUN fcis-29036 48 4 , , PUNCT fcis-29036 48 5 low low ADJ fcis-29036 48 6 cost cost NOUN fcis-29036 48 7 , , PUNCT fcis-29036 48 8 and and CCONJ fcis-29036 48 9 high high ADJ fcis-29036 48 10 performance performance NOUN fcis-29036 48 11 make make VERB fcis-29036 48 12 them they PRON fcis-29036 48 13 ideal ideal ADJ fcis-29036 48 14 for for ADP fcis-29036 48 15 on on ADP fcis-29036 48 16 - - PUNCT fcis-29036 48 17 site site NOUN fcis-29036 48 18 rapid rapid ADJ fcis-29036 48 19 detection detection NOUN fcis-29036 48 20 applications application NOUN fcis-29036 48 21 . . PUNCT fcis-29036 49 1 the the DET fcis-29036 49 2 rgb rgb PROPN fcis-29036 49 3 camera camera NOUN fcis-29036 49 4 is be AUX fcis-29036 49 5 equipped equip VERB fcis-29036 49 6 with with ADP fcis-29036 49 7 a a DET fcis-29036 49 8 wide wide ADJ fcis-29036 49 9 - - PUNCT fcis-29036 49 10 angle angle NOUN fcis-29036 49 11 lens lens NOUN fcis-29036 49 12 to to PART fcis-29036 49 13 capture capture VERB fcis-29036 49 14 image image NOUN fcis-29036 49 15 information information NOUN fcis-29036 49 16 over over ADP fcis-29036 49 17 a a DET fcis-29036 49 18 larger large ADJ fcis-29036 49 19 field field NOUN fcis-29036 49 20 of of ADP fcis-29036 49 21 view view NOUN fcis-29036 49 22 . . PUNCT fcis-29036 50 1 the the DET fcis-29036 50 2 system system NOUN fcis-29036 50 3 also also ADV fcis-29036 50 4 includes include VERB fcis-29036 50 5 four four NUM fcis-29036 50 6 50w 50w NOUN fcis-29036 50 7 halogen halogen NOUN fcis-29036 50 8 lamps lamp NOUN fcis-29036 50 9 as as ADP fcis-29036 50 10 the the DET fcis-29036 50 11 lighting lighting NOUN fcis-29036 50 12 source source NOUN fcis-29036 50 13 . . PUNCT fcis-29036 51 1 the the DET fcis-29036 51 2 conveyor conveyor NOUN fcis-29036 51 3 belt belt NOUN fcis-29036 51 4 , , PUNCT fcis-29036 51 5 made make VERB fcis-29036 51 6 of of ADP fcis-29036 51 7 white white ADJ fcis-29036 51 8 food food NOUN fcis-29036 51 9 - - PUNCT fcis-29036 51 10 grade grade NOUN fcis-29036 51 11 pu pu PROPN fcis-29036 51 12 material material NOUN fcis-29036 51 13 , , PUNCT fcis-29036 51 14 serves serve VERB fcis-29036 51 15 as as ADP fcis-29036 51 16 the the DET fcis-29036 51 17 background background NOUN fcis-29036 51 18 for for ADP fcis-29036 51 19 material material NOUN fcis-29036 51 20 transport transport NOUN fcis-29036 51 21 and and CCONJ fcis-29036 51 22 also also ADV fcis-29036 51 23 functions function NOUN fcis-29036 51 24 as as ADP fcis-29036 51 25 a a DET fcis-29036 51 26 reference reference NOUN fcis-29036 51 27 whiteboard whiteboard NOUN fcis-29036 51 28 for for ADP fcis-29036 51 29 spectral spectral ADJ fcis-29036 51 30 image image NOUN fcis-29036 51 31 data datum NOUN fcis-29036 51 32 calibration calibration NOUN fcis-29036 51 33 . . PUNCT fcis-29036 52 1 light light ADJ fcis-29036 52 2 source source NOUN fcis-29036 52 3 multispectral multispectral PROPN fcis-29036 52 4 camera camera NOUN fcis-29036 52 5 plc plc PROPN fcis-29036 52 6 system system PROPN fcis-29036 52 7 conveyor conveyor NOUN fcis-29036 52 8 belt belt NOUN fcis-29036 52 9 computer computer NOUN fcis-29036 52 10 rgb rgb PROPN fcis-29036 52 11 camera camera PROPN fcis-29036 52 12 fig fig NOUN fcis-29036 52 13 2 2 NUM fcis-29036 52 14 . . PUNCT fcis-29036 52 15 hardware hardware NOUN fcis-29036 52 16 architecture architecture NOUN fcis-29036 52 17 of of ADP fcis-29036 52 18 the the DET fcis-29036 52 19 sugar sugar NOUN fcis-29036 52 20 impurity impurity NOUN fcis-29036 52 21 detection detection NOUN fcis-29036 52 22 system system NOUN fcis-29036 52 23 7 7 NUM fcis-29036 52 24 the the DET fcis-29036 52 25 workflow workflow NOUN fcis-29036 52 26 of of ADP fcis-29036 52 27 the the DET fcis-29036 52 28 system system NOUN fcis-29036 52 29 is be AUX fcis-29036 52 30 shown show VERB fcis-29036 52 31 in in ADP fcis-29036 52 32 figure figure NOUN fcis-29036 52 33 3 3 NUM fcis-29036 52 34 . . PUNCT fcis-29036 52 35 initially initially ADV fcis-29036 52 36 , , PUNCT fcis-29036 52 37 the the DET fcis-29036 52 38 hardware hardware NOUN fcis-29036 52 39 setup setup NOUN fcis-29036 52 40 is be AUX fcis-29036 52 41 debugged debug VERB fcis-29036 52 42 and and CCONJ fcis-29036 52 43 adjusted adjust VERB fcis-29036 52 44 , , PUNCT fcis-29036 52 45 including include VERB fcis-29036 52 46 the the DET fcis-29036 52 47 calibration calibration NOUN fcis-29036 52 48 of of ADP fcis-29036 52 49 the the DET fcis-29036 52 50 cameras camera NOUN fcis-29036 52 51 and and CCONJ fcis-29036 52 52 light light NOUN fcis-29036 52 53 sources source NOUN fcis-29036 52 54 . . PUNCT fcis-29036 53 1 next next ADV fcis-29036 53 2 , , PUNCT fcis-29036 53 3 the the DET fcis-29036 53 4 whiteboard whiteboard NOUN fcis-29036 53 5 and and CCONJ fcis-29036 53 6 calibration calibration NOUN fcis-29036 53 7 board board NOUN fcis-29036 53 8 are be AUX fcis-29036 53 9 captured capture VERB fcis-29036 53 10 to to PART fcis-29036 53 11 obtain obtain VERB fcis-29036 53 12 the the DET fcis-29036 53 13 whiteboard whiteboard NOUN fcis-29036 53 14 data datum NOUN fcis-29036 53 15 and and CCONJ fcis-29036 53 16 the the DET fcis-29036 53 17 geometric geometric ADJ fcis-29036 53 18 parameters parameter NOUN fcis-29036 53 19 of of ADP fcis-29036 53 20 the the DET fcis-29036 53 21 camera camera NOUN fcis-29036 53 22 . . PUNCT fcis-29036 54 1 these these DET fcis-29036 54 2 data datum NOUN fcis-29036 54 3 are be AUX fcis-29036 54 4 used use VERB fcis-29036 54 5 during during ADP fcis-29036 54 6 the the DET fcis-29036 54 7 preprocessing preprocessing NOUN fcis-29036 54 8 stage stage NOUN fcis-29036 54 9 to to PART fcis-29036 54 10 calibrate calibrate VERB fcis-29036 54 11 the the DET fcis-29036 54 12 acquired acquire VERB fcis-29036 54 13 spectral spectral ADJ fcis-29036 54 14 images image NOUN fcis-29036 54 15 , , PUNCT fcis-29036 54 16 ensuring ensure VERB fcis-29036 54 17 both both DET fcis-29036 54 18 image image NOUN fcis-29036 54 19 quality quality NOUN fcis-29036 54 20 and and CCONJ fcis-29036 54 21 data datum NOUN fcis-29036 54 22 consistency consistency NOUN fcis-29036 54 23 . . PUNCT fcis-29036 55 1 after after ADP fcis-29036 55 2 preprocessing preprocesse VERB fcis-29036 55 3 , , PUNCT fcis-29036 55 4 the the DET fcis-29036 55 5 data datum NOUN fcis-29036 55 6 are be AUX fcis-29036 55 7 used use VERB fcis-29036 55 8 for for ADP fcis-29036 55 9 model model NOUN fcis-29036 55 10 training training NOUN fcis-29036 55 11 . . PUNCT fcis-29036 56 1 finally finally ADV fcis-29036 56 2 , , PUNCT fcis-29036 56 3 the the DET fcis-29036 56 4 results result NOUN fcis-29036 56 5 are be AUX fcis-29036 56 6 displayed display VERB fcis-29036 56 7 on on ADP fcis-29036 56 8 the the DET fcis-29036 56 9 software software NOUN fcis-29036 56 10 interface interface NOUN fcis-29036 56 11 , , PUNCT fcis-29036 56 12 showing show VERB fcis-29036 56 13 the the DET fcis-29036 56 14 processed process VERB fcis-29036 56 15 images image NOUN fcis-29036 56 16 . . PUNCT fcis-29036 57 1 white white PROPN fcis-29036 57 2     SPACE fcis-29036 57 3 board board NOUN fcis-29036 57 4     SPACE fcis-29036 57 5 and and CCONJ fcis-29036 57 6     SPACE fcis-29036 57 7 calibration calibration NOUN fcis-29036 57 8     SPACE fcis-29036 57 9 board board NOUN fcis-29036 57 10     SPACE fcis-29036 57 11 collection collection NOUN fcis-29036 57 12 data datum NOUN fcis-29036 57 13     SPACE fcis-29036 57 14 collection collection NOUN fcis-29036 57 15 model model NOUN fcis-29036 57 16     SPACE fcis-29036 57 17 processing processing NOUN fcis-29036 57 18 result result NOUN fcis-29036 57 19 camera camera NOUN fcis-29036 57 20 parameters parameter NOUN fcis-29036 57 21 calculated calculate VERB fcis-29036 57 22 from from ADP fcis-29036 57 23 whiteboard whiteboard NOUN fcis-29036 57 24 data datum NOUN fcis-29036 57 25 and and CCONJ fcis-29036 57 26 calibration calibration NOUN fcis-29036 57 27 board board NOUN fcis-29036 57 28 system system NOUN fcis-29036 57 29     SPACE fcis-29036 57 30 hardware hardware NOUN fcis-29036 57 31     SPACE fcis-29036 57 32 adjustment adjustment NOUN fcis-29036 57 33 preprocessing preprocesse VERB fcis-29036 57 34 fig fig NOUN fcis-29036 57 35 3 3 NUM fcis-29036 57 36 . . PUNCT fcis-29036 57 37 system system NOUN fcis-29036 57 38 workflow workflow NOUN fcis-29036 57 39 diagram diagram NOUN fcis-29036 57 40 3 3 NUM fcis-29036 57 41 . . PUNCT fcis-29036 57 42 data datum NOUN fcis-29036 57 43 preparation preparation NOUN fcis-29036 57 44 and and CCONJ fcis-29036 57 45 processing processing NOUN fcis-29036 57 46 3.1 3.1 NUM fcis-29036 57 47 . . PUNCT fcis-29036 58 1 experimental experimental ADJ fcis-29036 58 2 materials material NOUN fcis-29036 58 3 white white ADJ fcis-29036 58 4 sugar sugar NOUN fcis-29036 58 5 and and CCONJ fcis-29036 58 6 impurity impurity NOUN fcis-29036 58 7 samples sample NOUN fcis-29036 58 8 sourced source VERB fcis-29036 58 9 from from ADP fcis-29036 58 10 a a DET fcis-29036 58 11 leading lead VERB fcis-29036 58 12 domestic domestic ADJ fcis-29036 58 13 producer producer NOUN fcis-29036 58 14 were be AUX fcis-29036 58 15 used use VERB fcis-29036 58 16 to to PART fcis-29036 58 17 replicate replicate VERB fcis-29036 58 18 real real ADJ fcis-29036 58 19 - - PUNCT fcis-29036 58 20 world world NOUN fcis-29036 58 21 production production NOUN fcis-29036 58 22 conditions condition NOUN fcis-29036 58 23 . . PUNCT fcis-29036 59 1 impurities impurity NOUN fcis-29036 59 2 included include VERB fcis-29036 59 3 common common ADJ fcis-29036 59 4 foreign foreign ADJ fcis-29036 59 5 objects object NOUN fcis-29036 59 6 such such ADJ fcis-29036 59 7 as as ADP fcis-29036 59 8 yellow yellow ADJ fcis-29036 59 9 sugar sugar NOUN fcis-29036 59 10 , , PUNCT fcis-29036 59 11 hair hair NOUN fcis-29036 59 12 strands strand NOUN fcis-29036 59 13 , , PUNCT fcis-29036 59 14 rubber rubber NOUN fcis-29036 59 15 fragments fragment NOUN fcis-29036 59 16 , , PUNCT fcis-29036 59 17 packaging packaging NOUN fcis-29036 59 18 debris debris NOUN fcis-29036 59 19 , , PUNCT fcis-29036 59 20 black black ADJ fcis-29036 59 21 rubber rubber NOUN fcis-29036 59 22 , , PUNCT fcis-29036 59 23 blue blue ADJ fcis-29036 59 24 threads thread NOUN fcis-29036 59 25 , , PUNCT fcis-29036 59 26 and and CCONJ fcis-29036 59 27 metals metal NOUN fcis-29036 59 28 , , PUNCT fcis-29036 59 29 providing provide VERB fcis-29036 59 30 diverse diverse ADJ fcis-29036 59 31 and and CCONJ fcis-29036 59 32 challenging challenge VERB fcis-29036 59 33 scenarios scenario NOUN fcis-29036 59 34 for for ADP fcis-29036 59 35 testing testing NOUN fcis-29036 59 36 . . PUNCT fcis-29036 60 1 some some PRON fcis-29036 60 2 of of ADP fcis-29036 60 3 these these DET fcis-29036 60 4 impurities impurity NOUN fcis-29036 60 5 are be AUX fcis-29036 60 6 shown show VERB fcis-29036 60 7 in in ADP fcis-29036 60 8 figure figure NOUN fcis-29036 60 9 4 4 NUM fcis-29036 60 10 . . PUNCT fcis-29036 60 11 brown brown ADJ fcis-29036 60 12     SPACE fcis-29036 60 13 sugar sugar NOUN fcis-29036 60 14 rubber rubber NOUN fcis-29036 60 15 packaging packaging NOUN fcis-29036 60 16     SPACE fcis-29036 60 17 material material NOUN fcis-29036 60 18     SPACE fcis-29036 60 19 fragments fragment VERB fcis-29036 60 20 black black ADJ fcis-29036 60 21     SPACE fcis-29036 60 22 rubber rubber NOUN fcis-29036 60 23 blue blue ADJ fcis-29036 60 24     SPACE fcis-29036 60 25 thread thread NOUN fcis-29036 60 26 metal metal NOUN fcis-29036 60 27 hair hair NOUN fcis-29036 60 28 fig fig NOUN fcis-29036 60 29 4 4 NUM fcis-29036 60 30 . . PUNCT fcis-29036 60 31 color color NOUN fcis-29036 60 32 images image NOUN fcis-29036 60 33 of of ADP fcis-29036 60 34 different different ADJ fcis-29036 60 35 types type NOUN fcis-29036 60 36 of of ADP fcis-29036 60 37 impurities impurity NOUN fcis-29036 60 38 3.2 3.2 NUM fcis-29036 60 39 . . PUNCT fcis-29036 61 1 experimental experimental ADJ fcis-29036 61 2 data datum NOUN fcis-29036 61 3 preparation preparation NOUN fcis-29036 61 4 in in ADP fcis-29036 61 5 this this DET fcis-29036 61 6 experiment experiment NOUN fcis-29036 61 7 , , PUNCT fcis-29036 61 8 impurities impurity NOUN fcis-29036 61 9 were be AUX fcis-29036 61 10 randomly randomly ADV fcis-29036 61 11 selected select VERB fcis-29036 61 12 and and CCONJ fcis-29036 61 13 mixed mix VERB fcis-29036 61 14 with with ADP fcis-29036 61 15 white white ADJ fcis-29036 61 16 sugar sugar NOUN fcis-29036 61 17 to to PART fcis-29036 61 18 simulate simulate VERB fcis-29036 61 19 the the DET fcis-29036 61 20 randomness randomness NOUN fcis-29036 61 21 and and CCONJ fcis-29036 61 22 uncontrollability uncontrollability NOUN fcis-29036 61 23 of of ADP fcis-29036 61 24 impurities impurity NOUN fcis-29036 61 25 in in ADP fcis-29036 61 26 actual actual ADJ fcis-29036 61 27 production production NOUN fcis-29036 61 28 . . PUNCT fcis-29036 62 1 although although SCONJ fcis-29036 62 2 the the DET fcis-29036 62 3 impurity impurity NOUN fcis-29036 62 4 content content NOUN fcis-29036 62 5 in in ADP fcis-29036 62 6 real real ADJ fcis-29036 62 7 - - PUNCT fcis-29036 62 8 world world NOUN fcis-29036 62 9 production production NOUN fcis-29036 62 10 is be AUX fcis-29036 62 11 typically typically ADV fcis-29036 62 12 lower low ADJ fcis-29036 62 13 than than SCONJ fcis-29036 62 14 the the DET fcis-29036 62 15 settings setting NOUN fcis-29036 62 16 used use VERB fcis-29036 62 17 in in ADP fcis-29036 62 18 this this DET fcis-29036 62 19 experiment experiment NOUN fcis-29036 62 20 , , PUNCT fcis-29036 62 21 a a DET fcis-29036 62 22 higher high ADJ fcis-29036 62 23 impurity impurity NOUN fcis-29036 62 24 ratio ratio NOUN fcis-29036 62 25 was be AUX fcis-29036 62 26 intentionally intentionally ADV fcis-29036 62 27 chosen choose VERB fcis-29036 62 28 to to PART fcis-29036 62 29 ensure ensure VERB fcis-29036 62 30 the the DET fcis-29036 62 31 applicability applicability NOUN fcis-29036 62 32 of of ADP fcis-29036 62 33 the the DET fcis-29036 62 34 detection detection NOUN fcis-29036 62 35 method method NOUN fcis-29036 62 36 under under ADP fcis-29036 62 37 varying vary VERB fcis-29036 62 38 impurity impurity NOUN fcis-29036 62 39 content content NOUN fcis-29036 62 40 conditions condition NOUN fcis-29036 62 41 . . PUNCT fcis-29036 63 1 this this DET fcis-29036 63 2 design design NOUN fcis-29036 63 3 helps help VERB fcis-29036 63 4 to to PART fcis-29036 63 5 verify verify VERB fcis-29036 63 6 the the DET fcis-29036 63 7 robustness robustness NOUN fcis-29036 63 8 of of ADP fcis-29036 63 9 the the DET fcis-29036 63 10 proposed propose VERB fcis-29036 63 11 method method NOUN fcis-29036 63 12 in in ADP fcis-29036 63 13 complex complex ADJ fcis-29036 63 14 scenarios scenario NOUN fcis-29036 63 15 , , PUNCT fcis-29036 63 16 providing provide VERB fcis-29036 63 17 more more ADV fcis-29036 63 18 reliable reliable ADJ fcis-29036 63 19 technical technical ADJ fcis-29036 63 20 support support NOUN fcis-29036 63 21 and and CCONJ fcis-29036 63 22 assurance assurance NOUN fcis-29036 63 23 for for ADP fcis-29036 63 24 impurity impurity NOUN fcis-29036 63 25 detection detection NOUN fcis-29036 63 26 in in ADP fcis-29036 63 27 real real ADJ fcis-29036 63 28 - - PUNCT fcis-29036 63 29 world world NOUN fcis-29036 63 30 production production NOUN fcis-29036 63 31 environments environment NOUN fcis-29036 63 32 . . PUNCT fcis-29036 64 1 3.2.1 3.2.1 X fcis-29036 64 2 . . PUNCT fcis-29036 64 3 data datum NOUN fcis-29036 64 4 collection collection NOUN fcis-29036 64 5 before before ADP fcis-29036 64 6 data data NOUN fcis-29036 64 7 collection collection NOUN fcis-29036 64 8 , , PUNCT fcis-29036 64 9 the the DET fcis-29036 64 10 zoom zoom NOUN fcis-29036 64 11 ring ring NOUN fcis-29036 64 12 of of ADP fcis-29036 64 13 the the DET fcis-29036 64 14 rgb rgb PROPN fcis-29036 64 15 camera camera NOUN fcis-29036 64 16 's 's PART fcis-29036 64 17 wide wide ADJ fcis-29036 64 18 - - PUNCT fcis-29036 64 19 angle angle NOUN fcis-29036 64 20 lens lens NOUN fcis-29036 64 21 must must AUX fcis-29036 64 22 be be AUX fcis-29036 64 23 adjusted adjust VERB fcis-29036 64 24 to to PART fcis-29036 64 25 focus focus VERB fcis-29036 64 26 on on ADP fcis-29036 64 27 the the DET fcis-29036 64 28 plane plane NOUN fcis-29036 64 29 where where SCONJ fcis-29036 64 30 the the DET fcis-29036 64 31 samples sample NOUN fcis-29036 64 32 are be AUX fcis-29036 64 33 located locate VERB fcis-29036 64 34 . . PUNCT fcis-29036 65 1 additionally additionally ADV fcis-29036 65 2 , , PUNCT fcis-29036 65 3 the the DET fcis-29036 65 4 exposure exposure NOUN fcis-29036 65 5 time time NOUN fcis-29036 65 6 for for ADP fcis-29036 65 7 both both CCONJ fcis-29036 65 8 the the DET fcis-29036 65 9 multispectral multispectral ADJ fcis-29036 65 10 and and CCONJ fcis-29036 65 11 rgb rgb PROPN fcis-29036 65 12 cameras camera NOUN fcis-29036 65 13 needs need VERB fcis-29036 65 14 to to PART fcis-29036 65 15 be be AUX fcis-29036 65 16 finetuned finetune VERB fcis-29036 65 17 to to PART fcis-29036 65 18 avoid avoid VERB fcis-29036 65 19 overexposure overexposure NOUN fcis-29036 65 20 or or CCONJ fcis-29036 65 21 underexposure underexposure NOUN fcis-29036 65 22 , , PUNCT fcis-29036 65 23 which which PRON fcis-29036 65 24 could could AUX fcis-29036 65 25 affect affect VERB fcis-29036 65 26 the the DET fcis-29036 65 27 experimental experimental ADJ fcis-29036 65 28 results result NOUN fcis-29036 65 29 . . PUNCT fcis-29036 66 1 during during ADP fcis-29036 66 2 the the DET fcis-29036 66 3 formal formal ADJ fcis-29036 66 4 data data NOUN fcis-29036 66 5 collection collection NOUN fcis-29036 66 6 process process NOUN fcis-29036 66 7 , , PUNCT fcis-29036 66 8 the the DET fcis-29036 66 9 first first ADJ fcis-29036 66 10 step step NOUN fcis-29036 66 11 is be AUX fcis-29036 66 12 to to PART fcis-29036 66 13 capture capture VERB fcis-29036 66 14 the the DET fcis-29036 66 15 data datum NOUN fcis-29036 66 16 of of ADP fcis-29036 66 17 a a DET fcis-29036 66 18 white white ADJ fcis-29036 66 19 reference reference NOUN fcis-29036 66 20 board board NOUN fcis-29036 66 21 and and CCONJ fcis-29036 66 22 a a DET fcis-29036 66 23 calibration calibration NOUN fcis-29036 66 24 plate plate NOUN fcis-29036 66 25 for for ADP fcis-29036 66 26 system system NOUN fcis-29036 66 27 calibration calibration NOUN fcis-29036 66 28 . . PUNCT fcis-29036 67 1 this this PRON fcis-29036 67 2 ensures ensure VERB fcis-29036 67 3 the the DET fcis-29036 67 4 accuracy accuracy NOUN fcis-29036 67 5 and and CCONJ fcis-29036 67 6 consistency consistency NOUN fcis-29036 67 7 of of ADP fcis-29036 67 8 the the DET fcis-29036 67 9 spectral spectral ADJ fcis-29036 67 10 data datum NOUN fcis-29036 67 11 . . PUNCT fcis-29036 68 1 after after ADP fcis-29036 68 2 calibration calibration NOUN fcis-29036 68 3 , , PUNCT fcis-29036 68 4 white white ADJ fcis-29036 68 5 sugar sugar NOUN fcis-29036 68 6 is be AUX fcis-29036 68 7 mixed mix VERB fcis-29036 68 8 with with ADP fcis-29036 68 9 impurities impurity NOUN fcis-29036 68 10 , , PUNCT fcis-29036 68 11 and and CCONJ fcis-29036 68 12 data datum NOUN fcis-29036 68 13 collection collection NOUN fcis-29036 68 14 begins begin VERB fcis-29036 68 15 . . PUNCT fcis-29036 69 1 once once ADV fcis-29036 69 2 a a DET fcis-29036 69 3 set set NOUN fcis-29036 69 4 of of ADP fcis-29036 69 5 samples sample NOUN fcis-29036 69 6 is be AUX fcis-29036 69 7 photographed photograph VERB fcis-29036 69 8 , , PUNCT fcis-29036 69 9 the the DET fcis-29036 69 10 samples sample NOUN fcis-29036 69 11 must must AUX fcis-29036 69 12 be be AUX fcis-29036 69 13 promptly promptly ADV fcis-29036 69 14 replaced replace VERB fcis-29036 69 15 for for ADP fcis-29036 69 16 the the DET fcis-29036 69 17 next next ADJ fcis-29036 69 18 round round NOUN fcis-29036 69 19 of of ADP fcis-29036 69 20 data datum NOUN fcis-29036 69 21 collection collection NOUN fcis-29036 69 22 . . PUNCT fcis-29036 70 1 this this DET fcis-29036 70 2 process process NOUN fcis-29036 70 3 ensures ensure VERB fcis-29036 70 4 that that SCONJ fcis-29036 70 5 multiple multiple ADJ fcis-29036 70 6 different different ADJ fcis-29036 70 7 sample sample NOUN fcis-29036 70 8 combinations combination NOUN fcis-29036 70 9 are be AUX fcis-29036 70 10 captured capture VERB fcis-29036 70 11 , , PUNCT fcis-29036 70 12 improving improve VERB fcis-29036 70 13 the the DET fcis-29036 70 14 model model NOUN fcis-29036 70 15 's 's PART fcis-29036 70 16 generalization generalization NOUN fcis-29036 70 17 ability ability NOUN fcis-29036 70 18 and and CCONJ fcis-29036 70 19 detection detection NOUN fcis-29036 70 20 performance performance NOUN fcis-29036 70 21 . . PUNCT fcis-29036 71 1 3.2.2 3.2.2 X fcis-29036 71 2 . . PUNCT fcis-29036 71 3 spectral spectral ADJ fcis-29036 71 4 data datum NOUN fcis-29036 71 5 preprocessing preprocesse VERB fcis-29036 71 6 during during ADP fcis-29036 71 7 the the DET fcis-29036 71 8 imaging imaging NOUN fcis-29036 71 9 process process NOUN fcis-29036 71 10 , , PUNCT fcis-29036 71 11 spectral spectral ADJ fcis-29036 71 12 data datum NOUN fcis-29036 71 13 can can AUX fcis-29036 71 14 be be AUX fcis-29036 71 15 influenced influence VERB fcis-29036 71 16 by by ADP fcis-29036 71 17 various various ADJ fcis-29036 71 18 factors factor NOUN fcis-29036 71 19 , , PUNCT fcis-29036 71 20 making make VERB fcis-29036 71 21 data datum NOUN fcis-29036 71 22 preprocessing preprocesse VERB fcis-29036 71 23 particularly particularly ADV fcis-29036 71 24 important important ADJ fcis-29036 71 25 . . PUNCT fcis-29036 72 1 in in ADP fcis-29036 72 2 this this DET fcis-29036 72 3 study study NOUN fcis-29036 72 4 , , PUNCT fcis-29036 72 5 we we PRON fcis-29036 72 6 performed perform VERB fcis-29036 72 7 several several ADJ fcis-29036 72 8 preprocessing preprocessing NOUN fcis-29036 72 9 8 8 NUM fcis-29036 72 10 steps step NOUN fcis-29036 72 11 on on ADP fcis-29036 72 12 the the DET fcis-29036 72 13 spectral spectral ADJ fcis-29036 72 14 image image NOUN fcis-29036 72 15 data datum NOUN fcis-29036 72 16 , , PUNCT fcis-29036 72 17 including include VERB fcis-29036 72 18 cropping cropping NOUN fcis-29036 72 19 , , PUNCT fcis-29036 72 20 spectral spectral ADJ fcis-29036 72 21 calibration calibration NOUN fcis-29036 72 22 , , PUNCT fcis-29036 72 23 image image NOUN fcis-29036 72 24 registration registration NOUN fcis-29036 72 25 , , PUNCT fcis-29036 72 26 standard standard ADJ fcis-29036 72 27 normalization normalization NOUN fcis-29036 72 28 , , PUNCT fcis-29036 72 29 and and CCONJ fcis-29036 72 30 multiple multiple ADJ fcis-29036 72 31 scattering scatter VERB fcis-29036 72 32 correction correction NOUN fcis-29036 72 33 . . PUNCT fcis-29036 73 1 additionally additionally ADV fcis-29036 73 2 , , PUNCT fcis-29036 73 3 we we PRON fcis-29036 73 4 utilized utilize VERB fcis-29036 73 5 image image NOUN fcis-29036 73 6 fusion fusion NOUN fcis-29036 73 7 techniques technique NOUN fcis-29036 73 8 to to PART fcis-29036 73 9 combine combine VERB fcis-29036 73 10 high high ADJ fcis-29036 73 11 - - PUNCT fcis-29036 73 12 resolution resolution NOUN fcis-29036 73 13 images image NOUN fcis-29036 73 14 captured capture VERB fcis-29036 73 15 by by ADP fcis-29036 73 16 the the DET fcis-29036 73 17 rgb rgb PROPN fcis-29036 73 18 camera camera NOUN fcis-29036 73 19 with with ADP fcis-29036 73 20 spectral spectral ADJ fcis-29036 73 21 images image NOUN fcis-29036 73 22 obtained obtain VERB fcis-29036 73 23 from from ADP fcis-29036 73 24 the the DET fcis-29036 73 25 multispectral multispectral ADJ fcis-29036 73 26 camera camera NOUN fcis-29036 73 27 . . PUNCT fcis-29036 74 1 the the DET fcis-29036 74 2 processed process VERB fcis-29036 74 3 image image NOUN fcis-29036 74 4 data data NOUN fcis-29036 74 5 cube cube NOUN fcis-29036 74 6 contains contain VERB fcis-29036 74 7 13 13 NUM fcis-29036 74 8 channels channel NOUN fcis-29036 74 9 : : PUNCT fcis-29036 74 10 the the DET fcis-29036 74 11 first first ADJ fcis-29036 74 12 three three NUM fcis-29036 74 13 channels channel NOUN fcis-29036 74 14 represent represent VERB fcis-29036 74 15 the the DET fcis-29036 74 16 rgb rgb PROPN fcis-29036 74 17 image image NOUN fcis-29036 74 18 information information NOUN fcis-29036 74 19 , , PUNCT fcis-29036 74 20 while while SCONJ fcis-29036 74 21 the the DET fcis-29036 74 22 remaining remain VERB fcis-29036 74 23 10 10 NUM fcis-29036 74 24 channels channel NOUN fcis-29036 74 25 correspond correspond VERB fcis-29036 74 26 to to ADP fcis-29036 74 27 the the DET fcis-29036 74 28 multispectral multispectral ADJ fcis-29036 74 29 image image NOUN fcis-29036 74 30 data datum NOUN fcis-29036 74 31 at at ADP fcis-29036 74 32 10 10 NUM fcis-29036 74 33 different different ADJ fcis-29036 74 34 wavelengths wavelength NOUN fcis-29036 74 35 in in ADP fcis-29036 74 36 the the DET fcis-29036 74 37 near near ADV fcis-29036 74 38 - - PUNCT fcis-29036 74 39 infrared infrared ADJ fcis-29036 74 40 region region NOUN fcis-29036 74 41 . . PUNCT fcis-29036 75 1 these these DET fcis-29036 75 2 images image NOUN fcis-29036 75 3 provide provide VERB fcis-29036 75 4 a a DET fcis-29036 75 5 solid solid ADJ fcis-29036 75 6 foundation foundation NOUN fcis-29036 75 7 for for ADP fcis-29036 75 8 subsequent subsequent ADJ fcis-29036 75 9 analysis analysis NOUN fcis-29036 75 10 and and CCONJ fcis-29036 75 11 model model NOUN fcis-29036 75 12 training training NOUN fcis-29036 75 13 , , PUNCT fcis-29036 75 14 ensuring ensure VERB fcis-29036 75 15 that that SCONJ fcis-29036 75 16 the the DET fcis-29036 75 17 model model NOUN fcis-29036 75 18 can can AUX fcis-29036 75 19 effectively effectively ADV fcis-29036 75 20 learn learn VERB fcis-29036 75 21 and and CCONJ fcis-29036 75 22 improve improve VERB fcis-29036 75 23 detection detection NOUN fcis-29036 75 24 accuracy accuracy NOUN fcis-29036 75 25 . . PUNCT fcis-29036 76 1 spectral spectral ADJ fcis-29036 76 2 calibration calibration NOUN fcis-29036 76 3 is be AUX fcis-29036 76 4 performed perform VERB fcis-29036 76 5 to to PART fcis-29036 76 6 eliminate eliminate VERB fcis-29036 76 7 the the DET fcis-29036 76 8 effects effect NOUN fcis-29036 76 9 of of ADP fcis-29036 76 10 sensor sensor NOUN fcis-29036 76 11 baseline baseline NOUN fcis-29036 76 12 signals signal NOUN fcis-29036 76 13 and and CCONJ fcis-29036 76 14 lighting lighting NOUN fcis-29036 76 15 factors factor NOUN fcis-29036 76 16 , , PUNCT fcis-29036 76 17 ensuring ensure VERB fcis-29036 76 18 the the DET fcis-29036 76 19 accuracy accuracy NOUN fcis-29036 76 20 of of ADP fcis-29036 76 21 the the DET fcis-29036 76 22 spectral spectral ADJ fcis-29036 76 23 reflectance reflectance NOUN fcis-29036 76 24 in in ADP fcis-29036 76 25 the the DET fcis-29036 76 26 images image NOUN fcis-29036 76 27 . . PUNCT fcis-29036 77 1 after after ADP fcis-29036 77 2 each each DET fcis-29036 77 3 whiteboard whiteboard NOUN fcis-29036 77 4 data datum NOUN fcis-29036 77 5 collection collection NOUN fcis-29036 77 6 , , PUNCT fcis-29036 77 7 the the DET fcis-29036 77 8 camera camera NOUN fcis-29036 77 9 ’s ’s PART fcis-29036 77 10 built build VERB fcis-29036 77 11 - - PUNCT fcis-29036 77 12 in in ADP fcis-29036 77 13 program program NOUN fcis-29036 77 14 calibrates calibrate VERB fcis-29036 77 15 the the DET fcis-29036 77 16 images image NOUN fcis-29036 77 17 based base VERB fcis-29036 77 18 on on ADP fcis-29036 77 19 the the DET fcis-29036 77 20 following follow VERB fcis-29036 77 21 formula formula NOUN fcis-29036 77 22 ( ( PUNCT fcis-29036 77 23 1 1 NUM fcis-29036 77 24 ) ) PUNCT fcis-29036 77 25 to to PART fcis-29036 77 26 obtain obtain VERB fcis-29036 77 27 the the DET fcis-29036 77 28 corrected correct VERB fcis-29036 77 29 image image NOUN fcis-29036 77 30 : : PUNCT fcis-29036 77 31 reflenctance= reflenctance= ADJ fcis-29036 77 32 / / SYM fcis-29036 77 33 ( ( PUNCT fcis-29036 77 34 1 1 NUM fcis-29036 77 35 ) ) PUNCT fcis-29036 77 36 in in ADP fcis-29036 77 37 the the DET fcis-29036 77 38 above above ADJ fcis-29036 77 39 equation equation NOUN fcis-29036 77 40 , , PUNCT fcis-29036 77 41 reflectance reflectance NOUN fcis-29036 77 42 represents represent VERB fcis-29036 77 43 the the DET fcis-29036 77 44 corrected correct VERB fcis-29036 77 45 multispectral multispectral ADJ fcis-29036 77 46 data datum NOUN fcis-29036 77 47 , , PUNCT fcis-29036 77 48 rawdata rawdata PROPN fcis-29036 77 49 refers refer VERB fcis-29036 77 50 to to ADP fcis-29036 77 51 the the DET fcis-29036 77 52 raw raw ADJ fcis-29036 77 53 data datum NOUN fcis-29036 77 54 before before ADP fcis-29036 77 55 calibration calibration NOUN fcis-29036 77 56 , , PUNCT fcis-29036 77 57 and and CCONJ fcis-29036 77 58 whiteis whiteis VERB fcis-29036 77 59 the the DET fcis-29036 77 60 standard standard ADJ fcis-29036 77 61 reference reference NOUN fcis-29036 77 62 board board NOUN fcis-29036 77 63 data data PROPN fcis-29036 77 64 . . PUNCT fcis-29036 78 1 t1 t1 NOUN fcis-29036 78 2 and and CCONJ fcis-29036 78 3 t2 t2 NOUN fcis-29036 78 4 are be AUX fcis-29036 78 5 the the DET fcis-29036 78 6 exposure exposure NOUN fcis-29036 78 7 times time NOUN fcis-29036 78 8 for for ADP fcis-29036 78 9 capturing capture VERB fcis-29036 78 10 the the DET fcis-29036 78 11 sample sample NOUN fcis-29036 78 12 and and CCONJ fcis-29036 78 13 the the DET fcis-29036 78 14 whiteboard whiteboard NOUN fcis-29036 78 15 , , PUNCT fcis-29036 78 16 respectively respectively ADV fcis-29036 78 17 . . PUNCT fcis-29036 79 1 in in ADP fcis-29036 79 2 this this DET fcis-29036 79 3 study study NOUN fcis-29036 79 4 , , PUNCT fcis-29036 79 5 both both CCONJ fcis-29036 79 6 t1 t1 NOUN fcis-29036 79 7 and and CCONJ fcis-29036 79 8 t2 t2 NOUN fcis-29036 79 9 are be AUX fcis-29036 79 10 set set VERB fcis-29036 79 11 to to ADP fcis-29036 79 12 50 50 NUM fcis-29036 79 13 ms ms PROPN fcis-29036 79 14 . . PROPN fcis-29036 79 15 dark dark PROPN fcis-29036 79 16 and and CCONJ fcis-29036 79 17 dark dark ADJ fcis-29036 79 18 are be AUX fcis-29036 79 19 the the DET fcis-29036 79 20 dark dark ADJ fcis-29036 79 21 current current ADJ fcis-29036 79 22 values value NOUN fcis-29036 79 23 automatically automatically ADV fcis-29036 79 24 recorded record VERB fcis-29036 79 25 by by ADP fcis-29036 79 26 the the DET fcis-29036 79 27 camera camera NOUN fcis-29036 79 28 when when SCONJ fcis-29036 79 29 capturing capture VERB fcis-29036 79 30 the the DET fcis-29036 79 31 sample sample NOUN fcis-29036 79 32 and and CCONJ fcis-29036 79 33 the the DET fcis-29036 79 34 whiteboard whiteboard NOUN fcis-29036 79 35 . . PUNCT fcis-29036 80 1 the the DET fcis-29036 80 2 r r NOUN fcis-29036 80 3 is be AUX fcis-29036 80 4 the the DET fcis-29036 80 5 reflectance reflectance NOUN fcis-29036 80 6 of of ADP fcis-29036 80 7 the the DET fcis-29036 80 8 standard standard ADJ fcis-29036 80 9 whiteboard whiteboard NOUN fcis-29036 80 10 . . PUNCT fcis-29036 81 1 image image NOUN fcis-29036 81 2 registration registration NOUN fcis-29036 81 3 is be AUX fcis-29036 81 4 necessary necessary ADJ fcis-29036 81 5 because because SCONJ fcis-29036 81 6 different different ADJ fcis-29036 81 7 cameras camera NOUN fcis-29036 81 8 may may AUX fcis-29036 81 9 have have VERB fcis-29036 81 10 variations variation NOUN fcis-29036 81 11 in in ADP fcis-29036 81 12 viewpoint viewpoint NOUN fcis-29036 81 13 and and CCONJ fcis-29036 81 14 optical optical ADJ fcis-29036 81 15 characteristics characteristic NOUN fcis-29036 81 16 during during ADP fcis-29036 81 17 imaging imaging NOUN fcis-29036 81 18 , , PUNCT fcis-29036 81 19 which which PRON fcis-29036 81 20 results result VERB fcis-29036 81 21 in in ADP fcis-29036 81 22 inconsistencies inconsistency NOUN fcis-29036 81 23 in in ADP fcis-29036 81 24 the the DET fcis-29036 81 25 position position NOUN fcis-29036 81 26 , , PUNCT fcis-29036 81 27 shape shape NOUN fcis-29036 81 28 , , PUNCT fcis-29036 81 29 and and CCONJ fcis-29036 81 30 size size NOUN fcis-29036 81 31 of of ADP fcis-29036 81 32 the the DET fcis-29036 81 33 same same ADJ fcis-29036 81 34 object object NOUN fcis-29036 81 35 across across ADP fcis-29036 81 36 their their PRON fcis-29036 81 37 respective respective ADJ fcis-29036 81 38 images image NOUN fcis-29036 81 39 . . PUNCT fcis-29036 82 1 the the DET fcis-29036 82 2 process process NOUN fcis-29036 82 3 is be AUX fcis-29036 82 4 shown show VERB fcis-29036 82 5 in in ADP fcis-29036 82 6 figure figure NOUN fcis-29036 82 7 5 5 NUM fcis-29036 82 8 . . PUNCT fcis-29036 83 1 in in ADP fcis-29036 83 2 this this DET fcis-29036 83 3 study study NOUN fcis-29036 83 4 , , PUNCT fcis-29036 83 5 the the DET fcis-29036 83 6 aprilgrid aprilgrid PROPN fcis-29036 83 7 calibration calibration NOUN fcis-29036 83 8 board board NOUN fcis-29036 83 9 was be AUX fcis-29036 83 10 used use VERB fcis-29036 83 11 , , PUNCT fcis-29036 83 12 as as SCONJ fcis-29036 83 13 shown show VERB fcis-29036 83 14 in in ADP fcis-29036 83 15 figure figure NOUN fcis-29036 83 16 6 6 NUM fcis-29036 83 17 . . PUNCT fcis-29036 84 1 the the DET fcis-29036 84 2 aprilgrid aprilgrid PROPN fcis-29036 84 3 calibration calibration PROPN fcis-29036 84 4 board board NOUN fcis-29036 84 5 combines combine VERB fcis-29036 84 6 the the DET fcis-29036 84 7 features feature NOUN fcis-29036 84 8 of of ADP fcis-29036 84 9 traditional traditional ADJ fcis-29036 84 10 checkerboard checkerboard NOUN fcis-29036 84 11 grids grid NOUN fcis-29036 84 12 and and CCONJ fcis-29036 84 13 apriltag apriltag NOUN fcis-29036 84 14 markers marker NOUN fcis-29036 84 15 , , PUNCT fcis-29036 84 16 offering offer VERB fcis-29036 84 17 advantages advantage NOUN fcis-29036 84 18 such such ADJ fcis-29036 84 19 as as ADP fcis-29036 84 20 high high ADJ fcis-29036 84 21 - - PUNCT fcis-29036 84 22 precision precision NOUN fcis-29036 84 23 positioning positioning NOUN fcis-29036 84 24 , , PUNCT fcis-29036 84 25 strong strong ADJ fcis-29036 84 26 robustness robustness NOUN fcis-29036 84 27 , , PUNCT fcis-29036 84 28 ease ease NOUN fcis-29036 84 29 of of ADP fcis-29036 84 30 automation automation NOUN fcis-29036 84 31 , , PUNCT fcis-29036 84 32 and and CCONJ fcis-29036 84 33 applicability applicability NOUN fcis-29036 84 34 in in ADP fcis-29036 84 35 various various ADJ fcis-29036 84 36 scenarios scenario NOUN fcis-29036 84 37 . . PUNCT fcis-29036 85 1 rgb rgb PROPN fcis-29036 85 2     SPACE fcis-29036 85 3 camera camera NOUN fcis-29036 85 4     SPACE fcis-29036 85 5 and and CCONJ fcis-29036 85 6     SPACE fcis-29036 85 7 multispectral multispectral ADJ fcis-29036 85 8     SPACE fcis-29036 85 9 camera camera NOUN fcis-29036 85 10     SPACE fcis-29036 85 11 simultaneously simultaneously ADV fcis-29036 85 12     SPACE fcis-29036 85 13 shoot shoot VERB fcis-29036 85 14     SPACE fcis-29036 85 15 the the DET fcis-29036 85 16     SPACE fcis-29036 85 17 calibration calibration NOUN fcis-29036 85 18     SPACE fcis-29036 85 19 plate plate NOUN fcis-29036 85 20 extract extract NOUN fcis-29036 85 21     SPACE fcis-29036 85 22 feature feature NOUN fcis-29036 85 23     SPACE fcis-29036 85 24 points point NOUN fcis-29036 85 25     SPACE fcis-29036 85 26 separately separately ADV fcis-29036 85 27 matching match VERB fcis-29036 85 28     SPACE fcis-29036 85 29 feature feature NOUN fcis-29036 85 30     SPACE fcis-29036 85 31 points point NOUN fcis-29036 85 32 transformation transformation NOUN fcis-29036 85 33     SPACE fcis-29036 85 34 matrix matrix NOUN fcis-29036 85 35 image image NOUN fcis-29036 85 36     SPACE fcis-29036 85 37 alignment alignment NOUN fcis-29036 85 38 fig fig NOUN fcis-29036 85 39 5 5 NUM fcis-29036 85 40 . . PUNCT fcis-29036 85 41 image image NOUN fcis-29036 85 42 registration registration NOUN fcis-29036 85 43 workflow workflow NOUN fcis-29036 85 44 diagram diagram NOUN fcis-29036 85 45 fig fig NOUN fcis-29036 85 46 6 6 NUM fcis-29036 85 47 . . PUNCT fcis-29036 85 48 aprilgrid aprilgrid PROPN fcis-29036 85 49 calibration calibration PROPN fcis-29036 85 50 board board PROPN fcis-29036 85 51 3.2.3 3.2.3 NUM fcis-29036 85 52 . . PUNCT fcis-29036 86 1 spectral spectral ADJ fcis-29036 86 2 data datum NOUN fcis-29036 86 3 annotation annotation NOUN fcis-29036 86 4 after after SCONJ fcis-29036 86 5 the the DET fcis-29036 86 6 data data NOUN fcis-29036 86 7 preprocessing preprocessing NOUN fcis-29036 86 8 was be AUX fcis-29036 86 9 completed complete VERB fcis-29036 86 10 , , PUNCT fcis-29036 86 11 the the DET fcis-29036 86 12 images image NOUN fcis-29036 86 13 were be AUX fcis-29036 86 14 annotated annotate VERB fcis-29036 86 15 point point NOUN fcis-29036 86 16 by by ADP fcis-29036 86 17 point point NOUN fcis-29036 86 18 using use VERB fcis-29036 86 19 labelme labelme ADJ fcis-29036 86 20 software software NOUN fcis-29036 86 21 [ [ X fcis-29036 86 22 7 7 NUM fcis-29036 86 23 ] ] PUNCT fcis-29036 86 24 . . PUNCT fcis-29036 87 1 this this DET fcis-29036 87 2 tool tool NOUN fcis-29036 87 3 provides provide VERB fcis-29036 87 4 flexible flexible ADJ fcis-29036 87 5 and and CCONJ fcis-29036 87 6 efficient efficient ADJ fcis-29036 87 7 image image NOUN fcis-29036 87 8 annotation annotation NOUN fcis-29036 87 9 capabilities capability NOUN fcis-29036 87 10 , , PUNCT fcis-29036 87 11 helping help VERB fcis-29036 87 12 us we PRON fcis-29036 87 13 quickly quickly ADV fcis-29036 87 14 and and CCONJ fcis-29036 87 15 accurately accurately ADV fcis-29036 87 16 perform perform VERB fcis-29036 87 17 image image NOUN fcis-29036 87 18 segmentation segmentation NOUN fcis-29036 87 19 and and CCONJ fcis-29036 87 20 labeling labeling NOUN fcis-29036 87 21 . . PUNCT fcis-29036 88 1 using use VERB fcis-29036 88 2 labelme labelme NOUN fcis-29036 88 3 , , PUNCT fcis-29036 88 4 json json NOUN fcis-29036 88 5 files file NOUN fcis-29036 88 6 containing contain VERB fcis-29036 88 7 the the DET fcis-29036 88 8 coordinates coordinate NOUN fcis-29036 88 9 of of ADP fcis-29036 88 10 impurity impurity NOUN fcis-29036 88 11 regions region NOUN fcis-29036 88 12 were be AUX fcis-29036 88 13 generated generate VERB fcis-29036 88 14 . . PUNCT fcis-29036 89 1 additionally additionally ADV fcis-29036 89 2 , , PUNCT fcis-29036 89 3 scripts script NOUN fcis-29036 89 4 were be AUX fcis-29036 89 5 used use VERB fcis-29036 89 6 to to PART fcis-29036 89 7 process process VERB fcis-29036 89 8 these these DET fcis-29036 89 9 json json NOUN fcis-29036 89 10 files file NOUN fcis-29036 89 11 , , PUNCT fcis-29036 89 12 automatically automatically ADV fcis-29036 89 13 generating generate VERB fcis-29036 89 14 a a DET fcis-29036 89 15 visual visual ADJ fcis-29036 89 16 annotated annotate VERB fcis-29036 89 17 image image NOUN fcis-29036 89 18 ( ( PUNCT fcis-29036 89 19 visual.png visual.png NOUN fcis-29036 89 20 ) ) PUNCT fcis-29036 89 21 , , PUNCT fcis-29036 89 22 which which PRON fcis-29036 89 23 provided provide VERB fcis-29036 89 24 an an DET fcis-29036 89 25 intuitive intuitive ADJ fcis-29036 89 26 and and CCONJ fcis-29036 89 27 precise precise ADJ fcis-29036 89 28 reference reference NOUN fcis-29036 89 29 for for ADP fcis-29036 89 30 subsequent subsequent ADJ fcis-29036 89 31 model model NOUN fcis-29036 89 32 training training NOUN fcis-29036 89 33 . . PUNCT fcis-29036 90 1 4 4 X fcis-29036 90 2 . . X fcis-29036 90 3 introduction introduction NOUN fcis-29036 90 4 to to ADP fcis-29036 90 5 the the DET fcis-29036 90 6 attention attention NOUN fcis-29036 90 7 unet unet NOUN fcis-29036 90 8 model model NOUN fcis-29036 90 9 and and CCONJ fcis-29036 90 10 result result VERB fcis-29036 90 11 analysis analysis NOUN fcis-29036 90 12 u u NOUN fcis-29036 90 13 - - NOUN fcis-29036 90 14 net net ADJ fcis-29036 90 15 was be AUX fcis-29036 90 16 originally originally ADV fcis-29036 90 17 proposed propose VERB fcis-29036 90 18 for for ADP fcis-29036 90 19 segmenting segment VERB fcis-29036 90 20 neuronal neuronal ADJ fcis-29036 90 21 structures structure NOUN fcis-29036 90 22 by by ADP fcis-29036 90 23 training train VERB fcis-29036 90 24 an an DET fcis-29036 90 25 end end NOUN fcis-29036 90 26 - - PUNCT fcis-29036 90 27 to to ADP fcis-29036 90 28 - - PUNCT fcis-29036 90 29 end end VERB fcis-29036 90 30 fully fully ADV fcis-29036 90 31 convolutional convolutional ADJ fcis-29036 90 32 network network NOUN fcis-29036 90 33 on on ADP fcis-29036 90 34 electron electron NOUN fcis-29036 90 35 microscope microscope NOUN fcis-29036 90 36 images image NOUN fcis-29036 90 37 . . PUNCT fcis-29036 91 1 it it PRON fcis-29036 91 2 is be AUX fcis-29036 91 3 a a DET fcis-29036 91 4 classic classic ADJ fcis-29036 91 5 image image NOUN fcis-29036 91 6 segmentation segmentation NOUN fcis-29036 91 7 model model NOUN fcis-29036 91 8 . . PUNCT fcis-29036 92 1 with with ADP fcis-29036 92 2 its its PRON fcis-29036 92 3 encoder encoder NOUN fcis-29036 92 4 - - PUNCT fcis-29036 92 5 decoder decoder NOUN fcis-29036 92 6 architecture architecture NOUN fcis-29036 92 7 , , PUNCT fcis-29036 92 8 u u NOUN fcis-29036 92 9 - - NOUN fcis-29036 92 10 net net ADJ fcis-29036 92 11 is be AUX fcis-29036 92 12 capable capable ADJ fcis-29036 92 13 of of ADP fcis-29036 92 14 capturing capture VERB fcis-29036 92 15 both both CCONJ fcis-29036 92 16 global global ADJ fcis-29036 92 17 and and CCONJ fcis-29036 92 18 local local ADJ fcis-29036 92 19 information information NOUN fcis-29036 92 20 at at ADP fcis-29036 92 21 multiple multiple ADJ fcis-29036 92 22 scales scale NOUN fcis-29036 92 23 and and CCONJ fcis-29036 92 24 has have AUX fcis-29036 92 25 been be AUX fcis-29036 92 26 widely widely ADV fcis-29036 92 27 used use VERB fcis-29036 92 28 in in ADP fcis-29036 92 29 medical medical ADJ fcis-29036 92 30 image image NOUN fcis-29036 92 31 segmentation segmentation NOUN fcis-29036 92 32 and and CCONJ fcis-29036 92 33 natural natural ADJ fcis-29036 92 34 image image NOUN fcis-29036 92 35 segmentation segmentation NOUN fcis-29036 92 36 tasks task NOUN fcis-29036 93 1 [ [ X fcis-29036 93 2 8 8 NUM fcis-29036 93 3 ] ] PUNCT fcis-29036 93 4 . . PUNCT fcis-29036 94 1 9 9 NUM fcis-29036 94 2 attention attention NOUN fcis-29036 94 3 u u NOUN fcis-29036 94 4 - - NOUN fcis-29036 94 5 net net ADJ fcis-29036 94 6 is be AUX fcis-29036 94 7 an an DET fcis-29036 94 8 improved improved ADJ fcis-29036 94 9 version version NOUN fcis-29036 94 10 of of ADP fcis-29036 94 11 the the DET fcis-29036 94 12 classic classic ADJ fcis-29036 94 13 unet unet NOUN fcis-29036 94 14 that that PRON fcis-29036 94 15 incorporates incorporate VERB fcis-29036 94 16 an an DET fcis-29036 94 17 attention attention NOUN fcis-29036 94 18 mechanism mechanism NOUN fcis-29036 94 19 , , PUNCT fcis-29036 94 20 as as SCONJ fcis-29036 94 21 shown show VERB fcis-29036 94 22 in in ADP fcis-29036 94 23 figure figure NOUN fcis-29036 94 24 7 7 NUM fcis-29036 95 1 [ [ X fcis-29036 95 2 9 9 NUM fcis-29036 95 3 ] ] PUNCT fcis-29036 95 4 . . PUNCT fcis-29036 96 1 the the DET fcis-29036 96 2 attention attention NOUN fcis-29036 96 3 mechanism mechanism NOUN fcis-29036 96 4 enables enable VERB fcis-29036 96 5 the the DET fcis-29036 96 6 network network NOUN fcis-29036 96 7 to to PART fcis-29036 96 8 automatically automatically ADV fcis-29036 96 9 focus focus VERB fcis-29036 96 10 on on ADP fcis-29036 96 11 more more ADV fcis-29036 96 12 important important ADJ fcis-29036 96 13 feature feature NOUN fcis-29036 96 14 regions region NOUN fcis-29036 96 15 , , PUNCT fcis-29036 96 16 especially especially ADV fcis-29036 96 17 when when SCONJ fcis-29036 96 18 the the DET fcis-29036 96 19 contrast contrast NOUN fcis-29036 96 20 between between ADP fcis-29036 96 21 the the DET fcis-29036 96 22 target target NOUN fcis-29036 96 23 and and CCONJ fcis-29036 96 24 background background NOUN fcis-29036 96 25 is be AUX fcis-29036 96 26 low low ADJ fcis-29036 96 27 or or CCONJ fcis-29036 96 28 the the DET fcis-29036 96 29 boundaries boundary NOUN fcis-29036 96 30 are be AUX fcis-29036 96 31 unclear unclear ADJ fcis-29036 96 32 . . PUNCT fcis-29036 97 1 this this DET fcis-29036 97 2 mechanism mechanism NOUN fcis-29036 97 3 enhances enhance VERB fcis-29036 97 4 the the DET fcis-29036 97 5 ability ability NOUN fcis-29036 97 6 to to PART fcis-29036 97 7 capture capture VERB fcis-29036 97 8 critical critical ADJ fcis-29036 97 9 areas area NOUN fcis-29036 97 10 , , PUNCT fcis-29036 97 11 making make VERB fcis-29036 97 12 it it PRON fcis-29036 97 13 particularly particularly ADV fcis-29036 97 14 suitable suitable ADJ fcis-29036 97 15 for for ADP fcis-29036 97 16 organ organ NOUN fcis-29036 97 17 segmentation segmentation NOUN fcis-29036 97 18 in in ADP fcis-29036 97 19 medical medical ADJ fcis-29036 97 20 imaging imaging NOUN fcis-29036 97 21 [ [ X fcis-29036 97 22 10 10 NUM fcis-29036 97 23 ] ] PUNCT fcis-29036 97 24 . . PUNCT fcis-29036 98 1 in in ADP fcis-29036 98 2 the the DET fcis-29036 98 3 task task NOUN fcis-29036 98 4 of of ADP fcis-29036 98 5 white white ADJ fcis-29036 98 6 sugar sugar NOUN fcis-29036 98 7 impurity impurity NOUN fcis-29036 98 8 detection detection NOUN fcis-29036 98 9 , , PUNCT fcis-29036 98 10 the the DET fcis-29036 98 11 impurities impurity NOUN fcis-29036 98 12 are be AUX fcis-29036 98 13 typically typically ADV fcis-29036 98 14 small small ADJ fcis-29036 98 15 and and CCONJ fcis-29036 98 16 have have VERB fcis-29036 98 17 a a DET fcis-29036 98 18 low low ADJ fcis-29036 98 19 contrast contrast NOUN fcis-29036 98 20 with with ADP fcis-29036 98 21 the the DET fcis-29036 98 22 background background NOUN fcis-29036 98 23 ( ( PUNCT fcis-29036 98 24 white white ADJ fcis-29036 98 25 sugar sugar NOUN fcis-29036 98 26 ) ) PUNCT fcis-29036 98 27 . . PUNCT fcis-29036 99 1 the the DET fcis-29036 99 2 attention attention NOUN fcis-29036 99 3 mechanism mechanism NOUN fcis-29036 99 4 of of ADP fcis-29036 99 5 attention attention NOUN fcis-29036 99 6 u u NOUN fcis-29036 99 7 - - NOUN fcis-29036 99 8 net net ADJ fcis-29036 99 9 can can AUX fcis-29036 99 10 more more ADV fcis-29036 99 11 precisely precisely ADV fcis-29036 99 12 focus focus VERB fcis-29036 99 13 on on ADP fcis-29036 99 14 the the DET fcis-29036 99 15 impurity impurity NOUN fcis-29036 99 16 regions region NOUN fcis-29036 99 17 , , PUNCT fcis-29036 99 18 effectively effectively ADV fcis-29036 99 19 improving improve VERB fcis-29036 99 20 the the DET fcis-29036 99 21 accuracy accuracy NOUN fcis-29036 99 22 and and CCONJ fcis-29036 99 23 robustness robustness NOUN fcis-29036 99 24 of of ADP fcis-29036 99 25 detection detection NOUN fcis-29036 99 26 . . PUNCT fcis-29036 100 1 through through ADP fcis-29036 100 2 this this DET fcis-29036 100 3 mechanism mechanism NOUN fcis-29036 100 4 , , PUNCT fcis-29036 100 5 the the DET fcis-29036 100 6 network network NOUN fcis-29036 100 7 is be AUX fcis-29036 100 8 able able ADJ fcis-29036 100 9 to to PART fcis-29036 100 10 better well ADV fcis-29036 100 11 ignore ignore VERB fcis-29036 100 12 irrelevant irrelevant ADJ fcis-29036 100 13 background background NOUN fcis-29036 100 14 information information NOUN fcis-29036 100 15 and and CCONJ fcis-29036 100 16 highlight highlight VERB fcis-29036 100 17 the the DET fcis-29036 100 18 features feature NOUN fcis-29036 100 19 of of ADP fcis-29036 100 20 the the DET fcis-29036 100 21 impurities impurity NOUN fcis-29036 100 22 , , PUNCT fcis-29036 100 23 significantly significantly ADV fcis-29036 100 24 enhancing enhance VERB fcis-29036 100 25 segmentation segmentation NOUN fcis-29036 100 26 performance performance NOUN fcis-29036 100 27 . . PUNCT fcis-29036 101 1 this this PRON fcis-29036 101 2 makes make VERB fcis-29036 101 3 attention attention NOUN fcis-29036 101 4 u u NOUN fcis-29036 101 5 - - NOUN fcis-29036 101 6 net net ADJ fcis-29036 101 7 more more ADV fcis-29036 101 8 effective effective ADJ fcis-29036 101 9 in in ADP fcis-29036 101 10 recognizing recognize VERB fcis-29036 101 11 and and CCONJ fcis-29036 101 12 locating locate VERB fcis-29036 101 13 impurities impurity NOUN fcis-29036 101 14 in in ADP fcis-29036 101 15 complex complex ADJ fcis-29036 101 16 scenarios scenario NOUN fcis-29036 101 17 , , PUNCT fcis-29036 101 18 providing provide VERB fcis-29036 101 19 reliable reliable ADJ fcis-29036 101 20 support support NOUN fcis-29036 101 21 for for ADP fcis-29036 101 22 automated automate VERB fcis-29036 101 23 detection detection NOUN fcis-29036 101 24 . . PUNCT fcis-29036 102 1 in in ADP fcis-29036 102 2 the the DET fcis-29036 102 3 white white PROPN fcis-29036 102 4 sugar sugar PROPN fcis-29036 102 5 impurity impurity NOUN fcis-29036 102 6 detection detection NOUN fcis-29036 102 7 task task NOUN fcis-29036 102 8 of of ADP fcis-29036 102 9 this this DET fcis-29036 102 10 study study NOUN fcis-29036 102 11 , , PUNCT fcis-29036 102 12 due due ADP fcis-29036 102 13 to to ADP fcis-29036 102 14 the the DET fcis-29036 102 15 introduction introduction NOUN fcis-29036 102 16 of of ADP fcis-29036 102 17 a a DET fcis-29036 102 18 ten ten NUM fcis-29036 102 19 - - PUNCT fcis-29036 102 20 band band NOUN fcis-29036 102 21 multispectral multispectral ADJ fcis-29036 102 22 camera camera NOUN fcis-29036 102 23 , , PUNCT fcis-29036 102 24 the the DET fcis-29036 102 25 model model NOUN fcis-29036 102 26 was be AUX fcis-29036 102 27 modified modify VERB fcis-29036 102 28 to to PART fcis-29036 102 29 handle handle VERB fcis-29036 102 30 a a DET fcis-29036 102 31 thirteen thirteen NUM fcis-29036 102 32 - - PUNCT fcis-29036 102 33 channel channel NOUN fcis-29036 102 34 multispectral multispectral ADJ fcis-29036 102 35 image image NOUN fcis-29036 102 36 input input NOUN fcis-29036 102 37 , , PUNCT fcis-29036 102 38 instead instead ADV fcis-29036 102 39 of of ADP fcis-29036 102 40 the the DET fcis-29036 102 41 traditional traditional ADJ fcis-29036 102 42 singlechannel singlechannel NOUN fcis-29036 102 43 grayscale grayscale NOUN fcis-29036 102 44 or or CCONJ fcis-29036 102 45 three three NUM fcis-29036 102 46 - - PUNCT fcis-29036 102 47 channel channel NOUN fcis-29036 102 48 rgb rgb PROPN fcis-29036 102 49 images image NOUN fcis-29036 102 50 . . PUNCT fcis-29036 103 1 this this DET fcis-29036 103 2 modification modification NOUN fcis-29036 103 3 allows allow VERB fcis-29036 103 4 the the DET fcis-29036 103 5 model model NOUN fcis-29036 103 6 to to PART fcis-29036 103 7 fully fully ADV fcis-29036 103 8 utilize utilize VERB fcis-29036 103 9 the the DET fcis-29036 103 10 multispectral multispectral ADJ fcis-29036 103 11 image image NOUN fcis-29036 103 12 data datum NOUN fcis-29036 103 13 , , PUNCT fcis-29036 103 14 enhancing enhance VERB fcis-29036 103 15 its its PRON fcis-29036 103 16 ability ability NOUN fcis-29036 103 17 to to PART fcis-29036 103 18 detect detect VERB fcis-29036 103 19 and and CCONJ fcis-29036 103 20 identify identify VERB fcis-29036 103 21 impurities impurity NOUN fcis-29036 103 22 in in ADP fcis-29036 103 23 white white ADJ fcis-29036 103 24 sugar sugar NOUN fcis-29036 103 25 more more ADV fcis-29036 103 26 effectively effectively ADV fcis-29036 103 27 . . PUNCT fcis-29036 104 1 fig fig NOUN fcis-29036 104 2 7 7 NUM fcis-29036 104 3 . . PUNCT fcis-29036 104 4 attention attention NOUN fcis-29036 104 5 unet unet NOUN fcis-29036 104 6 model model NOUN fcis-29036 104 7 structure structure NOUN fcis-29036 104 8 4.1 4.1 NUM fcis-29036 104 9 . . PUNCT fcis-29036 105 1 model model NOUN fcis-29036 105 2 evaluation evaluation NOUN fcis-29036 105 3 metrics metric NOUN fcis-29036 105 4 to to PART fcis-29036 105 5 comprehensively comprehensively ADV fcis-29036 105 6 evaluate evaluate VERB fcis-29036 105 7 the the DET fcis-29036 105 8 model model NOUN fcis-29036 105 9 's 's PART fcis-29036 105 10 performance performance NOUN fcis-29036 105 11 in in ADP fcis-29036 105 12 the the DET fcis-29036 105 13 sugar sugar NOUN fcis-29036 105 14 impurity impurity NOUN fcis-29036 105 15 detection detection NOUN fcis-29036 105 16 task task NOUN fcis-29036 105 17 , , PUNCT fcis-29036 105 18 this this DET fcis-29036 105 19 study study NOUN fcis-29036 105 20 uses use VERB fcis-29036 105 21 several several ADJ fcis-29036 105 22 common common ADJ fcis-29036 105 23 segmentation segmentation NOUN fcis-29036 105 24 evaluation evaluation NOUN fcis-29036 105 25 metrics metric NOUN fcis-29036 105 26 , , PUNCT fcis-29036 105 27 including include VERB fcis-29036 105 28 accuracy accuracy NOUN fcis-29036 105 29 , , PUNCT fcis-29036 105 30 mean mean ADJ fcis-29036 105 31 pixel pixel PROPN fcis-29036 105 32 accuracy accuracy NOUN fcis-29036 105 33 ( ( PUNCT fcis-29036 105 34 mpa mpa PROPN fcis-29036 105 35 ) ) PUNCT fcis-29036 105 36 , , PUNCT fcis-29036 106 1 mean mean VERB fcis-29036 106 2 intersection intersection NOUN fcis-29036 106 3 over over ADP fcis-29036 106 4 union union NOUN fcis-29036 106 5 ( ( PUNCT fcis-29036 106 6 miou miou NOUN fcis-29036 106 7 ) ) PUNCT fcis-29036 106 8 , , PUNCT fcis-29036 106 9 and and CCONJ fcis-29036 106 10 dice dice VERB fcis-29036 106 11 coefficient coefficient NOUN fcis-29036 106 12 . . PUNCT fcis-29036 107 1 4.2 4.2 NUM fcis-29036 107 2 . . PUNCT fcis-29036 108 1 analysis analysis NOUN fcis-29036 108 2 of of ADP fcis-29036 108 3 model model ADJ fcis-29036 108 4 comparison comparison NOUN fcis-29036 108 5 results result NOUN fcis-29036 108 6 during during ADP fcis-29036 108 7 the the DET fcis-29036 108 8 training training NOUN fcis-29036 108 9 process process NOUN fcis-29036 108 10 , , PUNCT fcis-29036 108 11 we we PRON fcis-29036 108 12 recorded record VERB fcis-29036 108 13 the the DET fcis-29036 108 14 model model NOUN fcis-29036 108 15 's 's PART fcis-29036 108 16 loss loss NOUN fcis-29036 108 17 and and CCONJ fcis-29036 108 18 dice dice NOUN fcis-29036 108 19 coefficient coefficient NOUN fcis-29036 108 20 , , PUNCT fcis-29036 108 21 as as SCONJ fcis-29036 108 22 shown show VERB fcis-29036 108 23 in in ADP fcis-29036 108 24 figure figure NOUN fcis-29036 108 25 8 8 NUM fcis-29036 108 26 , , PUNCT fcis-29036 108 27 to to PART fcis-29036 108 28 evaluate evaluate VERB fcis-29036 108 29 the the DET fcis-29036 108 30 model model NOUN fcis-29036 108 31 's 's PART fcis-29036 108 32 performance performance NOUN fcis-29036 108 33 . . PUNCT fcis-29036 109 1 these these DET fcis-29036 109 2 metrics metric NOUN fcis-29036 109 3 provide provide VERB fcis-29036 109 4 us we PRON fcis-29036 109 5 with with ADP fcis-29036 109 6 an an DET fcis-29036 109 7 intuitive intuitive ADJ fcis-29036 109 8 understanding understanding NOUN fcis-29036 109 9 of of ADP fcis-29036 109 10 the the DET fcis-29036 109 11 model model NOUN fcis-29036 109 12 's 's PART fcis-29036 109 13 training training NOUN fcis-29036 109 14 effectiveness effectiveness NOUN fcis-29036 109 15 and and CCONJ fcis-29036 109 16 help help NOUN fcis-29036 109 17 in in ADP fcis-29036 109 18 further far ADV fcis-29036 109 19 optimizing optimize VERB fcis-29036 109 20 the the DET fcis-29036 109 21 model model NOUN fcis-29036 109 22 's 's PART fcis-29036 109 23 parameters parameter NOUN fcis-29036 109 24 and and CCONJ fcis-29036 109 25 training training NOUN fcis-29036 109 26 strategy strategy NOUN fcis-29036 109 27 . . PUNCT fcis-29036 110 1 ( ( PUNCT fcis-29036 110 2 a a X fcis-29036 110 3 ) ) PUNCT fcis-29036 110 4 ( ( PUNCT fcis-29036 110 5 b b X fcis-29036 110 6 ) ) PUNCT fcis-29036 110 7 fig fig NOUN fcis-29036 110 8 8 8 NUM fcis-29036 110 9 . . PUNCT fcis-29036 111 1 ( ( PUNCT fcis-29036 111 2 a a X fcis-29036 111 3 ) ) PUNCT fcis-29036 111 4 loss loss NOUN fcis-29036 111 5 during during ADP fcis-29036 111 6 the the DET fcis-29036 111 7 training training NOUN fcis-29036 111 8 process process NOUN fcis-29036 111 9 and and CCONJ fcis-29036 111 10 ( ( PUNCT fcis-29036 111 11 b b NOUN fcis-29036 111 12 ) ) PUNCT fcis-29036 111 13 real real ADJ fcis-29036 111 14 - - PUNCT fcis-29036 111 15 time time NOUN fcis-29036 111 16 changes change NOUN fcis-29036 111 17 in in ADP fcis-29036 111 18 the the DET fcis-29036 111 19 dice dice NOUN fcis-29036 111 20 coefficient coefficient NOUN fcis-29036 111 21 as as SCONJ fcis-29036 111 22 seen see VERB fcis-29036 111 23 from from ADP fcis-29036 111 24 the the DET fcis-29036 111 25 figure figure NOUN fcis-29036 111 26 above above ADV fcis-29036 111 27 , , PUNCT fcis-29036 111 28 the the DET fcis-29036 111 29 loss loss NOUN fcis-29036 111 30 value value NOUN fcis-29036 111 31 gradually gradually ADV fcis-29036 111 32 decreases decrease VERB fcis-29036 111 33 during during ADP fcis-29036 111 34 the the DET fcis-29036 111 35 training training NOUN fcis-29036 111 36 process process NOUN fcis-29036 111 37 , , PUNCT fcis-29036 111 38 while while SCONJ fcis-29036 111 39 the the DET fcis-29036 111 40 dice dice NOUN fcis-29036 111 41 coefficient coefficient NOUN fcis-29036 111 42 steadily steadily ADV fcis-29036 111 43 increases increase VERB fcis-29036 111 44 , , PUNCT fcis-29036 111 45 indicating indicate VERB fcis-29036 111 46 that that SCONJ fcis-29036 111 47 the the DET fcis-29036 111 48 model model NOUN fcis-29036 111 49 exhibits exhibit VERB fcis-29036 111 50 good good ADJ fcis-29036 111 51 convergence convergence NOUN fcis-29036 111 52 and and CCONJ fcis-29036 111 53 improving improve VERB fcis-29036 111 54 segmentation segmentation NOUN fcis-29036 111 55 performance performance NOUN fcis-29036 111 56 . . PUNCT fcis-29036 112 1 these these DET fcis-29036 112 2 results result NOUN fcis-29036 112 3 suggest suggest VERB fcis-29036 112 4 that that SCONJ fcis-29036 112 5 after after ADP fcis-29036 112 6 multiple multiple ADJ fcis-29036 112 7 training training NOUN fcis-29036 112 8 rounds round NOUN fcis-29036 112 9 , , PUNCT fcis-29036 112 10 the the DET fcis-29036 112 11 model model NOUN fcis-29036 112 12 can can AUX fcis-29036 112 13 effectively effectively ADV fcis-29036 112 14 adapt adapt VERB fcis-29036 112 15 to to ADP fcis-29036 112 16 the the DET fcis-29036 112 17 data datum NOUN fcis-29036 112 18 and and CCONJ fcis-29036 112 19 enhance enhance VERB fcis-29036 112 20 its its PRON fcis-29036 112 21 ability ability NOUN fcis-29036 112 22 to to PART fcis-29036 112 23 detect detect VERB fcis-29036 112 24 and and CCONJ fcis-29036 112 25 segment segment NOUN fcis-29036 112 26 impurities impurity NOUN fcis-29036 112 27 in in ADP fcis-29036 112 28 white white ADJ fcis-29036 112 29 sugar sugar NOUN fcis-29036 112 30 . . PUNCT fcis-29036 113 1 10 10 NUM fcis-29036 113 2 in in ADP fcis-29036 113 3 this this DET fcis-29036 113 4 experiment experiment NOUN fcis-29036 113 5 , , PUNCT fcis-29036 113 6 we we PRON fcis-29036 113 7 optimized optimize VERB fcis-29036 113 8 the the DET fcis-29036 113 9 attention attention NOUN fcis-29036 113 10 unet unet NOUN fcis-29036 113 11 model model NOUN fcis-29036 113 12 . . PUNCT fcis-29036 114 1 traditional traditional ADJ fcis-29036 114 2 attention attention NOUN fcis-29036 114 3 unet unet NOUN fcis-29036 114 4 typically typically ADV fcis-29036 114 5 uses use VERB fcis-29036 114 6 a a DET fcis-29036 114 7 single single ADJ fcis-29036 114 8 rgb rgb PROPN fcis-29036 114 9 image image NOUN fcis-29036 114 10 as as ADP fcis-29036 114 11 input input NOUN fcis-29036 114 12 , , PUNCT fcis-29036 114 13 whereas whereas SCONJ fcis-29036 114 14 in in ADP fcis-29036 114 15 this this DET fcis-29036 114 16 experiment experiment NOUN fcis-29036 114 17 , , PUNCT fcis-29036 114 18 we we PRON fcis-29036 114 19 modified modify VERB fcis-29036 114 20 the the DET fcis-29036 114 21 input input NOUN fcis-29036 114 22 channels channel NOUN fcis-29036 114 23 to to ADP fcis-29036 114 24 thirteen thirteen NUM fcis-29036 114 25 images image NOUN fcis-29036 114 26 to to PART fcis-29036 114 27 better well ADV fcis-29036 114 28 suit suit VERB fcis-29036 114 29 our our PRON fcis-29036 114 30 dataset dataset NOUN fcis-29036 114 31 . . PUNCT fcis-29036 115 1 during during ADP fcis-29036 115 2 the the DET fcis-29036 115 3 training training NOUN fcis-29036 115 4 process process NOUN fcis-29036 115 5 , , PUNCT fcis-29036 115 6 the the DET fcis-29036 115 7 dataset dataset NOUN fcis-29036 115 8 was be AUX fcis-29036 115 9 divided divide VERB fcis-29036 115 10 into into ADP fcis-29036 115 11 training training NOUN fcis-29036 115 12 , , PUNCT fcis-29036 115 13 validation validation NOUN fcis-29036 115 14 , , PUNCT fcis-29036 115 15 and and CCONJ fcis-29036 115 16 test test NOUN fcis-29036 115 17 sets set NOUN fcis-29036 115 18 in in ADP fcis-29036 115 19 a a DET fcis-29036 115 20 60 60 NUM fcis-29036 115 21 % % NOUN fcis-29036 115 22 , , PUNCT fcis-29036 115 23 20 20 NUM fcis-29036 115 24 % % NOUN fcis-29036 115 25 , , PUNCT fcis-29036 115 26 and and CCONJ fcis-29036 115 27 20 20 NUM fcis-29036 115 28 % % NOUN fcis-29036 115 29 ratio ratio NOUN fcis-29036 115 30 , , PUNCT fcis-29036 115 31 respectively respectively ADV fcis-29036 115 32 , , PUNCT fcis-29036 115 33 for for ADP fcis-29036 115 34 model model NOUN fcis-29036 115 35 training training NOUN fcis-29036 115 36 , , PUNCT fcis-29036 115 37 parameter parameter NOUN fcis-29036 115 38 tuning tuning NOUN fcis-29036 115 39 , , PUNCT fcis-29036 115 40 and and CCONJ fcis-29036 115 41 performance performance NOUN fcis-29036 115 42 evaluation evaluation NOUN fcis-29036 115 43 . . PUNCT fcis-29036 116 1 additionally additionally ADV fcis-29036 116 2 , , PUNCT fcis-29036 116 3 we we PRON fcis-29036 116 4 performed perform VERB fcis-29036 116 5 data datum NOUN fcis-29036 116 6 processing processing NOUN fcis-29036 116 7 for for ADP fcis-29036 116 8 both both DET fcis-29036 116 9 types type NOUN fcis-29036 116 10 of of ADP fcis-29036 116 11 input input NOUN fcis-29036 116 12 and and CCONJ fcis-29036 116 13 conducted conduct VERB fcis-29036 116 14 a a DET fcis-29036 116 15 detailed detailed ADJ fcis-29036 116 16 analysis analysis NOUN fcis-29036 116 17 of of ADP fcis-29036 116 18 the the DET fcis-29036 116 19 results result NOUN fcis-29036 116 20 for for ADP fcis-29036 116 21 each each DET fcis-29036 116 22 category category NOUN fcis-29036 116 23 as as ADV fcis-29036 116 24 well well ADV fcis-29036 116 25 as as ADP fcis-29036 116 26 the the DET fcis-29036 116 27 overall overall ADJ fcis-29036 116 28 results result NOUN fcis-29036 116 29 , , PUNCT fcis-29036 116 30 as as SCONJ fcis-29036 116 31 shown show VERB fcis-29036 116 32 in in ADP fcis-29036 116 33 table table NOUN fcis-29036 116 34 1 1 NUM fcis-29036 116 35 . . PUNCT fcis-29036 116 36 table table NOUN fcis-29036 116 37 1 1 NUM fcis-29036 116 38 . . PUNCT fcis-29036 116 39 recognition recognition NOUN fcis-29036 116 40 results result NOUN fcis-29036 116 41 of of ADP fcis-29036 116 42 different different ADJ fcis-29036 116 43 categories category NOUN fcis-29036 116 44 under under ADP fcis-29036 116 45 models model NOUN fcis-29036 116 46 with with ADP fcis-29036 116 47 different different ADJ fcis-29036 116 48 input input NOUN fcis-29036 116 49 image image NOUN fcis-29036 116 50 numbers number NOUN fcis-29036 116 51 ( ( PUNCT fcis-29036 116 52 all all PRON fcis-29036 116 53 refers refer VERB fcis-29036 116 54 to to ADP fcis-29036 116 55 the the DET fcis-29036 116 56 overall overall ADJ fcis-29036 116 57 results result NOUN fcis-29036 116 58 of of ADP fcis-29036 116 59 the the DET fcis-29036 116 60 model model NOUN fcis-29036 116 61 for for ADP fcis-29036 116 62 sugar sugar NOUN fcis-29036 116 63 and and CCONJ fcis-29036 116 64 impurities impurity NOUN fcis-29036 116 65 ) ) PUNCT fcis-29036 116 66 input input NOUN fcis-29036 116 67 image image NOUN fcis-29036 116 68 category category NOUN fcis-29036 116 69 category category NOUN fcis-29036 116 70 mean mean VERB fcis-29036 116 71 intersection intersection NOUN fcis-29036 116 72 over over ADP fcis-29036 116 73 union union NOUN fcis-29036 116 74 ( ( PUNCT fcis-29036 116 75 miou miou NOUN fcis-29036 116 76 ) ) PUNCT fcis-29036 116 77 % % NOUN fcis-29036 117 1 mean mean NOUN fcis-29036 117 2 pixel pixel PROPN fcis-29036 117 3 accuracy accuracy NOUN fcis-29036 117 4 ( ( PUNCT fcis-29036 117 5 mpa mpa NOUN fcis-29036 117 6 ) ) PUNCT fcis-29036 117 7 % % NOUN fcis-29036 118 1 accuracy accuracy NOUN fcis-29036 118 2 % % NOUN fcis-29036 118 3 1 1 NUM fcis-29036 118 4 rgb rgb PROPN fcis-29036 118 5 image image NOUN fcis-29036 118 6 background background NOUN fcis-29036 118 7 99.66 99.66 NUM fcis-29036 118 8 99.67 99.67 NUM fcis-29036 118 9 99.80 99.80 NUM fcis-29036 118 10 impurities impurity NOUN fcis-29036 118 11 31.22 31.22 NUM fcis-29036 118 12 94.16 94.16 NUM fcis-29036 118 13 31.84 31.84 NUM fcis-29036 118 14 all all PRON fcis-29036 118 15 66.44 66.44 NUM fcis-29036 118 16 96.91 96.91 NUM fcis-29036 118 17 99.66 99.66 NUM fcis-29036 118 18 r r NOUN fcis-29036 118 19 , , PUNCT fcis-29036 118 20 g g PROPN fcis-29036 118 21 , , PUNCT fcis-29036 118 22 b b NOUN fcis-29036 118 23 , , PUNCT fcis-29036 118 24 and and CCONJ fcis-29036 118 25 ten ten NUM fcis-29036 118 26 multispectral multispectral ADJ fcis-29036 118 27 images image NOUN fcis-29036 118 28 background background NOUN fcis-29036 118 29 99.97 99.97 NUM fcis-29036 118 30 99.97 99.97 NUM fcis-29036 118 31 99.97 99.97 NUM fcis-29036 118 32 impurities impurity NOUN fcis-29036 118 33 95.42 95.42 NUM fcis-29036 118 34 87.50 87.50 NUM fcis-29036 118 35 95.17 95.17 NUM fcis-29036 118 36 all all DET fcis-29036 118 37 92.70 92.70 NUM fcis-29036 118 38 93.74 93.74 NUM fcis-29036 118 39 99.97 99.97 NUM fcis-29036 118 40 from from ADP fcis-29036 118 41 the the DET fcis-29036 118 42 table table NOUN fcis-29036 118 43 above above ADV fcis-29036 118 44 , , PUNCT fcis-29036 118 45 it it PRON fcis-29036 118 46 can can AUX fcis-29036 118 47 be be AUX fcis-29036 118 48 observed observe VERB fcis-29036 118 49 that that SCONJ fcis-29036 118 50 when when SCONJ fcis-29036 118 51 the the DET fcis-29036 118 52 input input NOUN fcis-29036 118 53 is be AUX fcis-29036 118 54 a a DET fcis-29036 118 55 single single ADJ fcis-29036 118 56 rgb rgb PROPN fcis-29036 118 57 image image NOUN fcis-29036 118 58 , , PUNCT fcis-29036 118 59 the the DET fcis-29036 118 60 model model NOUN fcis-29036 118 61 performs perform VERB fcis-29036 118 62 excellently excellently ADV fcis-29036 118 63 in in ADP fcis-29036 118 64 the the DET fcis-29036 118 65 background background NOUN fcis-29036 118 66 ( ( PUNCT fcis-29036 118 67 i.e. i.e. X fcis-29036 118 68 , , PUNCT fcis-29036 118 69 white white ADJ fcis-29036 118 70 sugar sugar NOUN fcis-29036 118 71 ) ) PUNCT fcis-29036 118 72 , , PUNCT fcis-29036 118 73 with with ADP fcis-29036 118 74 the the DET fcis-29036 118 75 mean mean ADJ fcis-29036 118 76 intersection intersection NOUN fcis-29036 118 77 over over ADP fcis-29036 118 78 union union NOUN fcis-29036 118 79 ( ( PUNCT fcis-29036 118 80 miou miou NOUN fcis-29036 118 81 ) ) PUNCT fcis-29036 118 82 , , PUNCT fcis-29036 118 83 mean mean VERB fcis-29036 118 84 pixel pixel PROPN fcis-29036 118 85 accuracy accuracy NOUN fcis-29036 118 86 ( ( PUNCT fcis-29036 118 87 mpa mpa PROPN fcis-29036 118 88 ) ) PUNCT fcis-29036 118 89 , , PUNCT fcis-29036 118 90 and and CCONJ fcis-29036 118 91 overall overall ADJ fcis-29036 118 92 accuracy accuracy NOUN fcis-29036 118 93 all all PRON fcis-29036 118 94 exceeding exceed VERB fcis-29036 118 95 99 99 NUM fcis-29036 118 96 % % NOUN fcis-29036 118 97 . . PUNCT fcis-29036 119 1 however however ADV fcis-29036 119 2 , , PUNCT fcis-29036 119 3 the the DET fcis-29036 119 4 model model NOUN fcis-29036 119 5 's 's PART fcis-29036 119 6 performance performance NOUN fcis-29036 119 7 in in ADP fcis-29036 119 8 impurity impurity NOUN fcis-29036 119 9 detection detection NOUN fcis-29036 119 10 is be AUX fcis-29036 119 11 unsatisfactory unsatisfactory ADJ fcis-29036 119 12 , , PUNCT fcis-29036 119 13 with with ADP fcis-29036 119 14 both both CCONJ fcis-29036 119 15 the the DET fcis-29036 119 16 mean mean ADJ fcis-29036 119 17 intersection intersection NOUN fcis-29036 119 18 over over ADP fcis-29036 119 19 union union NOUN fcis-29036 119 20 and and CCONJ fcis-29036 119 21 accuracy accuracy NOUN fcis-29036 119 22 falling fall VERB fcis-29036 119 23 below below ADP fcis-29036 119 24 50 50 NUM fcis-29036 119 25 % % NOUN fcis-29036 119 26 . . PUNCT fcis-29036 120 1 despite despite SCONJ fcis-29036 120 2 the the DET fcis-29036 120 3 high high ADJ fcis-29036 120 4 overall overall ADJ fcis-29036 120 5 accuracy accuracy NOUN fcis-29036 120 6 , , PUNCT fcis-29036 120 7 this this PRON fcis-29036 120 8 is be AUX fcis-29036 120 9 primarily primarily ADV fcis-29036 120 10 due due ADJ fcis-29036 120 11 to to ADP fcis-29036 120 12 the the DET fcis-29036 120 13 much much ADV fcis-29036 120 14 higher high ADJ fcis-29036 120 15 proportion proportion NOUN fcis-29036 120 16 of of ADP fcis-29036 120 17 white white ADJ fcis-29036 120 18 sugar sugar NOUN fcis-29036 120 19 compared compare VERB fcis-29036 120 20 to to ADP fcis-29036 120 21 impurities impurity NOUN fcis-29036 120 22 , , PUNCT fcis-29036 120 23 which which PRON fcis-29036 120 24 skews skew VERB fcis-29036 120 25 the the DET fcis-29036 120 26 overall overall ADJ fcis-29036 120 27 accuracy accuracy NOUN fcis-29036 120 28 . . PUNCT fcis-29036 121 1 in in ADP fcis-29036 121 2 contrast contrast NOUN fcis-29036 121 3 , , PUNCT fcis-29036 121 4 the the DET fcis-29036 121 5 improved improved ADJ fcis-29036 121 6 attention attention NOUN fcis-29036 121 7 unet unet NOUN fcis-29036 121 8 model model NOUN fcis-29036 121 9 shows show VERB fcis-29036 121 10 significant significant ADJ fcis-29036 121 11 improvement improvement NOUN fcis-29036 121 12 in in ADP fcis-29036 121 13 both both DET fcis-29036 121 14 background background NOUN fcis-29036 121 15 and and CCONJ fcis-29036 121 16 impurity impurity NOUN fcis-29036 121 17 detection detection NOUN fcis-29036 121 18 . . PUNCT fcis-29036 122 1 particularly particularly ADV fcis-29036 122 2 in in ADP fcis-29036 122 3 impurity impurity NOUN fcis-29036 122 4 detection detection NOUN fcis-29036 122 5 , , PUNCT fcis-29036 122 6 the the DET fcis-29036 122 7 mean mean ADJ fcis-29036 122 8 intersection intersection NOUN fcis-29036 122 9 over over ADP fcis-29036 122 10 union union NOUN fcis-29036 122 11 ( ( PUNCT fcis-29036 122 12 miou miou NOUN fcis-29036 122 13 ) ) PUNCT fcis-29036 122 14 increased increase VERB fcis-29036 122 15 to to ADP fcis-29036 122 16 95.42 95.42 NUM fcis-29036 122 17 % % NOUN fcis-29036 122 18 , , PUNCT fcis-29036 122 19 and and CCONJ fcis-29036 122 20 the the DET fcis-29036 122 21 overall overall ADJ fcis-29036 122 22 accuracy accuracy NOUN fcis-29036 122 23 reached reach VERB fcis-29036 122 24 95.17 95.17 NUM fcis-29036 122 25 % % NOUN fcis-29036 122 26 . . PUNCT fcis-29036 123 1 this this PRON fcis-29036 123 2 indicates indicate VERB fcis-29036 123 3 that that SCONJ fcis-29036 123 4 the the DET fcis-29036 123 5 introduction introduction NOUN fcis-29036 123 6 of of ADP fcis-29036 123 7 multispectral multispectral ADJ fcis-29036 123 8 images image NOUN fcis-29036 123 9 significantly significantly ADV fcis-29036 123 10 enhanced enhance VERB fcis-29036 123 11 the the DET fcis-29036 123 12 model model NOUN fcis-29036 123 13 's 's PART fcis-29036 123 14 detection detection NOUN fcis-29036 123 15 capability capability NOUN fcis-29036 123 16 in in ADP fcis-29036 123 17 complex complex ADJ fcis-29036 123 18 scenarios scenario NOUN fcis-29036 123 19 , , PUNCT fcis-29036 123 20 improving improve VERB fcis-29036 123 21 its its PRON fcis-29036 123 22 ability ability NOUN fcis-29036 123 23 to to PART fcis-29036 123 24 identify identify VERB fcis-29036 123 25 and and CCONJ fcis-29036 123 26 segment segment NOUN fcis-29036 123 27 impurities impurity NOUN fcis-29036 123 28 in in ADP fcis-29036 123 29 white white ADJ fcis-29036 123 30 sugar sugar NOUN fcis-29036 123 31 . . PUNCT fcis-29036 124 1 4.3 4.3 NUM fcis-29036 124 2 . . PUNCT fcis-29036 124 3 attention attention NOUN fcis-29036 124 4 u_net u_net NOUN fcis-29036 124 5 training training NOUN fcis-29036 124 6 results result NOUN fcis-29036 124 7 and and CCONJ fcis-29036 124 8 visualization visualization NOUN fcis-29036 124 9 results result NOUN fcis-29036 124 10 in in ADP fcis-29036 124 11 this this DET fcis-29036 124 12 test test NOUN fcis-29036 124 13 , , PUNCT fcis-29036 124 14 we we PRON fcis-29036 124 15 not not PART fcis-29036 124 16 only only ADV fcis-29036 124 17 evaluated evaluate VERB fcis-29036 124 18 the the DET fcis-29036 124 19 model model NOUN fcis-29036 124 20 's 's PART fcis-29036 124 21 ability ability NOUN fcis-29036 124 22 to to PART fcis-29036 124 23 recognize recognize VERB fcis-29036 124 24 the the DET fcis-29036 124 25 impurity impurity NOUN fcis-29036 124 26 types type NOUN fcis-29036 124 27 present present ADJ fcis-29036 124 28 in in ADP fcis-29036 124 29 the the DET fcis-29036 124 30 training training NOUN fcis-29036 124 31 dataset dataset NOUN fcis-29036 124 32 but but CCONJ fcis-29036 124 33 also also ADV fcis-29036 124 34 specifically specifically ADV fcis-29036 124 35 examined examine VERB fcis-29036 124 36 its its PRON fcis-29036 124 37 performance performance NOUN fcis-29036 124 38 in in ADP fcis-29036 124 39 detecting detect VERB fcis-29036 124 40 new new ADJ fcis-29036 124 41 impurities impurity NOUN fcis-29036 124 42 ( ( PUNCT fcis-29036 124 43 i.e. i.e. X fcis-29036 124 44 , , PUNCT fcis-29036 124 45 impurities impurity NOUN fcis-29036 124 46 that that PRON fcis-29036 124 47 were be AUX fcis-29036 124 48 not not PART fcis-29036 124 49 present present ADJ fcis-29036 124 50 during during ADP fcis-29036 124 51 the the DET fcis-29036 124 52 training training NOUN fcis-29036 124 53 phase phase NOUN fcis-29036 124 54 ) ) PUNCT fcis-29036 124 55 . . PUNCT fcis-29036 125 1 the the DET fcis-29036 125 2 results result NOUN fcis-29036 125 3 for for ADP fcis-29036 125 4 the the DET fcis-29036 125 5 recognition recognition NOUN fcis-29036 125 6 of of ADP fcis-29036 125 7 known know VERB fcis-29036 125 8 impurities impurity NOUN fcis-29036 125 9 are be AUX fcis-29036 125 10 shown show VERB fcis-29036 125 11 in in ADP fcis-29036 125 12 figure figure NOUN fcis-29036 125 13 9 9 NUM fcis-29036 125 14 , , PUNCT fcis-29036 125 15 and and CCONJ fcis-29036 125 16 the the DET fcis-29036 125 17 results result NOUN fcis-29036 125 18 for for ADP fcis-29036 125 19 the the DET fcis-29036 125 20 detection detection NOUN fcis-29036 125 21 of of ADP fcis-29036 125 22 new new ADJ fcis-29036 125 23 impurities impurity NOUN fcis-29036 125 24 are be AUX fcis-29036 125 25 shown show VERB fcis-29036 125 26 in in ADP fcis-29036 125 27 figure figure NOUN fcis-29036 125 28 10 10 NUM fcis-29036 125 29 . . PUNCT fcis-29036 126 1 ( ( PUNCT fcis-29036 126 2 a a X fcis-29036 126 3 ) ) PUNCT fcis-29036 126 4 ( ( PUNCT fcis-29036 126 5 b b X fcis-29036 126 6 ) ) PUNCT fcis-29036 126 7 1 1 NUM fcis-29036 126 8 cm cm NUM fcis-29036 126 9 1 1 NUM fcis-29036 126 10 cm cm NOUN fcis-29036 126 11 fig fig NOUN fcis-29036 126 12 9 9 NUM fcis-29036 126 13 . . PUNCT fcis-29036 127 1 ( ( PUNCT fcis-29036 127 2 a a X fcis-29036 127 3 ) ) PUNCT fcis-29036 127 4 original original ADJ fcis-29036 127 5 image image NOUN fcis-29036 127 6 of of ADP fcis-29036 127 7 impurities impurity NOUN fcis-29036 127 8 from from ADP fcis-29036 127 9 existing exist VERB fcis-29036 127 10 categories category NOUN fcis-29036 127 11 in in ADP fcis-29036 127 12 the the DET fcis-29036 127 13 training training NOUN fcis-29036 127 14 set set NOUN fcis-29036 127 15 ; ; PUNCT fcis-29036 127 16 ( ( PUNCT fcis-29036 127 17 b b X fcis-29036 127 18 ) ) PUNCT fcis-29036 127 19 predicted predict VERB fcis-29036 127 20 result result NOUN fcis-29036 127 21 image image NOUN fcis-29036 127 22 as as SCONJ fcis-29036 127 23 shown show VERB fcis-29036 127 24 in in ADP fcis-29036 127 25 the the DET fcis-29036 127 26 above above ADJ fcis-29036 127 27 figures figure NOUN fcis-29036 127 28 , , PUNCT fcis-29036 127 29 figure figure NOUN fcis-29036 127 30 9(a 9(a NUM fcis-29036 127 31 ) ) PUNCT fcis-29036 127 32 displays display VERB fcis-29036 127 33 the the DET fcis-29036 127 34 original original ADJ fcis-29036 127 35 multispectral multispectral ADJ fcis-29036 127 36 image image NOUN fcis-29036 127 37 , , PUNCT fcis-29036 127 38 with with ADP fcis-29036 127 39 the the DET fcis-29036 127 40 impurity impurity NOUN fcis-29036 127 41 locations location NOUN fcis-29036 127 42 marked mark VERB fcis-29036 127 43 by by ADP fcis-29036 127 44 red red ADJ fcis-29036 127 45 circles circle NOUN fcis-29036 127 46 . . PUNCT fcis-29036 128 1 figure figure NOUN fcis-29036 128 2 9(b 9(b NUM fcis-29036 128 3 ) ) PUNCT fcis-29036 128 4 presents present VERB fcis-29036 128 5 the the DET fcis-29036 128 6 model model NOUN fcis-29036 128 7 's 's PART fcis-29036 128 8 prediction prediction NOUN fcis-29036 128 9 results result NOUN fcis-29036 128 10 for for ADP fcis-29036 128 11 the the DET fcis-29036 128 12 image image NOUN fcis-29036 128 13 in in ADP fcis-29036 128 14 figure figure NOUN fcis-29036 128 15 9(a 9(a NUM fcis-29036 128 16 ) ) PUNCT fcis-29036 128 17 . . PUNCT fcis-29036 129 1 a a DET fcis-29036 129 2 comparison comparison NOUN fcis-29036 129 3 reveals reveal VERB fcis-29036 129 4 that that SCONJ fcis-29036 129 5 most most ADJ fcis-29036 129 6 impurities impurity NOUN fcis-29036 129 7 are be AUX fcis-29036 129 8 accurately accurately ADV fcis-29036 129 9 identified identify VERB fcis-29036 129 10 , , PUNCT fcis-29036 129 11 and and CCONJ fcis-29036 129 12 the the DET fcis-29036 129 13 model model NOUN fcis-29036 129 14 performs perform VERB fcis-29036 129 15 well well ADV fcis-29036 129 16 in in ADP fcis-29036 129 17 distinguishing distinguish VERB fcis-29036 129 18 between between ADP fcis-29036 129 19 sugar sugar NOUN fcis-29036 129 20 and and CCONJ fcis-29036 129 21 impurities impurity NOUN fcis-29036 129 22 . . PUNCT fcis-29036 130 1 as as SCONJ fcis-29036 130 2 shown show VERB fcis-29036 130 3 in in ADP fcis-29036 130 4 the the DET fcis-29036 130 5 above above ADJ fcis-29036 130 6 figure figure NOUN fcis-29036 130 7 , , PUNCT fcis-29036 130 8 the the DET fcis-29036 130 9 detection detection NOUN fcis-29036 130 10 of of ADP fcis-29036 130 11 previously previously ADV fcis-29036 130 12 unseen unseen ADJ fcis-29036 130 13 impurities impurity NOUN fcis-29036 130 14 during during ADP fcis-29036 130 15 the the DET fcis-29036 130 16 model model NOUN fcis-29036 130 17 training training NOUN fcis-29036 130 18 process process NOUN fcis-29036 130 19 is be AUX fcis-29036 130 20 demonstrated demonstrate VERB fcis-29036 130 21 . . PUNCT fcis-29036 131 1 figure figure VERB fcis-29036 131 2 10(a 10(a NUM fcis-29036 131 3 ) ) PUNCT fcis-29036 131 4 displays display VERB fcis-29036 131 5 the the DET fcis-29036 131 6 original original ADJ fcis-29036 131 7 image image NOUN fcis-29036 131 8 , , PUNCT fcis-29036 131 9 with with ADP fcis-29036 131 10 red red ADJ fcis-29036 131 11 circles circle NOUN fcis-29036 131 12 marking mark VERB fcis-29036 131 13 the the DET fcis-29036 131 14 newly newly ADV fcis-29036 131 15 introduced introduce VERB fcis-29036 131 16 impurities impurity NOUN fcis-29036 131 17 and and CCONJ fcis-29036 131 18 cyan cyan NOUN fcis-29036 131 19 circles circle NOUN fcis-29036 131 20 indicating indicate VERB fcis-29036 131 21 the the DET fcis-29036 131 22 impurities impurity NOUN fcis-29036 131 23 that that PRON fcis-29036 131 24 were be AUX fcis-29036 131 25 present present ADJ fcis-29036 131 26 during during ADP fcis-29036 131 27 training training NOUN fcis-29036 131 28 . . PUNCT fcis-29036 132 1 figure figure NOUN fcis-29036 132 2 10(b 10(b NUM fcis-29036 132 3 ) ) PUNCT fcis-29036 132 4 shows show VERB fcis-29036 132 5 the the DET fcis-29036 132 6 model model NOUN fcis-29036 132 7 's 's PART fcis-29036 132 8 detection detection NOUN fcis-29036 132 9 results result NOUN fcis-29036 132 10 for for ADP fcis-29036 132 11 the the DET fcis-29036 132 12 image image NOUN fcis-29036 132 13 in in ADP fcis-29036 132 14 figure figure NOUN fcis-29036 132 15 10(a 10(a NUM fcis-29036 132 16 ) ) PUNCT fcis-29036 132 17 , , PUNCT fcis-29036 132 18 with with ADP fcis-29036 132 19 red red ADJ fcis-29036 132 20 circles circle NOUN fcis-29036 132 21 indicating indicate VERB fcis-29036 132 22 the the DET fcis-29036 132 23 impurities impurity NOUN fcis-29036 132 24 that that PRON fcis-29036 132 25 were be AUX fcis-29036 132 26 not not PART fcis-29036 132 27 detected detect VERB fcis-29036 132 28 and and CCONJ fcis-29036 132 29 black black ADJ fcis-29036 132 30 markings marking NOUN fcis-29036 132 31 representing represent VERB fcis-29036 132 32 correctly correctly ADV fcis-29036 132 33 identified identify VERB fcis-29036 132 34 impurities impurity NOUN fcis-29036 132 35 . . PUNCT fcis-29036 133 1 it it PRON fcis-29036 133 2 is be AUX fcis-29036 133 3 important important ADJ fcis-29036 133 4 to to PART fcis-29036 133 5 note note VERB fcis-29036 133 6 that that SCONJ fcis-29036 133 7 the the DET fcis-29036 133 8 undetected undetected ADJ fcis-29036 133 9 impurity impurity NOUN fcis-29036 133 10 is be AUX fcis-29036 133 11 a a DET fcis-29036 133 12 transparent transparent ADJ fcis-29036 133 13 film film NOUN fcis-29036 133 14 , , PUNCT fcis-29036 133 15 which which PRON fcis-29036 133 16 has have VERB fcis-29036 133 17 a a DET fcis-29036 133 18 high high ADJ fcis-29036 133 19 transmittance transmittance NOUN fcis-29036 133 20 for for ADP fcis-29036 133 21 incident incident NOUN fcis-29036 133 22 light light NOUN fcis-29036 133 23 and and CCONJ fcis-29036 133 24 a a DET fcis-29036 133 25 weak weak ADJ fcis-29036 133 26 reflected reflected ADJ fcis-29036 133 27 spectral spectral ADJ fcis-29036 133 28 signal signal NOUN fcis-29036 133 29 , , PUNCT fcis-29036 133 30 making make VERB fcis-29036 133 31 it it PRON fcis-29036 133 32 difficult difficult ADJ fcis-29036 133 33 for for SCONJ fcis-29036 133 34 the the DET fcis-29036 133 35 model model NOUN fcis-29036 133 36 to to PART fcis-29036 133 37 accurately accurately ADV fcis-29036 133 38 recognize recognize VERB fcis-29036 133 39 these these DET fcis-29036 133 40 impurities impurity NOUN fcis-29036 133 41 during during ADP fcis-29036 133 42 detection detection NOUN fcis-29036 133 43 . . PUNCT fcis-29036 134 1 11 11 NUM fcis-29036 135 1 the the DET fcis-29036 135 2 detection detection NOUN fcis-29036 135 3 results result NOUN fcis-29036 135 4 indicate indicate VERB fcis-29036 135 5 that that SCONJ fcis-29036 135 6 the the DET fcis-29036 135 7 model model NOUN fcis-29036 135 8 performs perform VERB fcis-29036 135 9 well well ADV fcis-29036 135 10 in in ADP fcis-29036 135 11 the the DET fcis-29036 135 12 practical practical ADJ fcis-29036 135 13 application application NOUN fcis-29036 135 14 of of ADP fcis-29036 135 15 white white ADJ fcis-29036 135 16 sugar sugar NOUN fcis-29036 135 17 impurity impurity NOUN fcis-29036 135 18 detection detection NOUN fcis-29036 135 19 , , PUNCT fcis-29036 135 20 achieving achieve VERB fcis-29036 135 21 efficient efficient ADJ fcis-29036 135 22 and and CCONJ fcis-29036 135 23 accurate accurate ADJ fcis-29036 135 24 impurity impurity NOUN fcis-29036 135 25 identification identification NOUN fcis-29036 135 26 . . PUNCT fcis-29036 136 1 additionally additionally ADV fcis-29036 136 2 , , PUNCT fcis-29036 136 3 the the DET fcis-29036 136 4 model model NOUN fcis-29036 136 5 also also ADV fcis-29036 136 6 demonstrates demonstrate VERB fcis-29036 136 7 strong strong ADJ fcis-29036 136 8 performance performance NOUN fcis-29036 136 9 when when SCONJ fcis-29036 136 10 recognizing recognize VERB fcis-29036 136 11 unfamiliar unfamiliar ADJ fcis-29036 136 12 impurity impurity NOUN fcis-29036 136 13 categories category NOUN fcis-29036 136 14 . . PUNCT fcis-29036 137 1 ( ( PUNCT fcis-29036 137 2 a a X fcis-29036 137 3 ) ) PUNCT fcis-29036 137 4 ( ( PUNCT fcis-29036 137 5 b b NOUN fcis-29036 137 6 ) ) PUNCT fcis-29036 137 7 1cm1 1cm1 NUM fcis-29036 137 8 cm cm NOUN fcis-29036 137 9 fig fig NOUN fcis-29036 137 10 10 10 NUM fcis-29036 137 11 . . PUNCT fcis-29036 138 1 ( ( PUNCT fcis-29036 138 2 a a X fcis-29036 138 3 ) ) PUNCT fcis-29036 138 4 original original ADJ fcis-29036 138 5 image image NOUN fcis-29036 138 6 with with ADP fcis-29036 138 7 added add VERB fcis-29036 138 8 new new ADJ fcis-29036 138 9 impurities impurity NOUN fcis-29036 138 10 ; ; PUNCT fcis-29036 138 11 ( ( PUNCT fcis-29036 138 12 b b X fcis-29036 138 13 ) ) PUNCT fcis-29036 138 14 model model NOUN fcis-29036 138 15 prediction prediction NOUN fcis-29036 138 16 result result NOUN fcis-29036 138 17 image image NOUN fcis-29036 138 18 5 5 NUM fcis-29036 138 19 . . PUNCT fcis-29036 138 20 summarize summarize VERB fcis-29036 138 21 in in ADP fcis-29036 138 22 the the DET fcis-29036 138 23 sugar sugar NOUN fcis-29036 138 24 processing processing NOUN fcis-29036 138 25 industry industry NOUN fcis-29036 138 26 , , PUNCT fcis-29036 138 27 existing exist VERB fcis-29036 138 28 impurity impurity NOUN fcis-29036 138 29 removal removal NOUN fcis-29036 138 30 techniques technique NOUN fcis-29036 138 31 fail fail VERB fcis-29036 138 32 to to PART fcis-29036 138 33 meet meet VERB fcis-29036 138 34 the the DET fcis-29036 138 35 demands demand NOUN fcis-29036 138 36 of of ADP fcis-29036 138 37 current current ADJ fcis-29036 138 38 production production NOUN fcis-29036 138 39 processes process NOUN fcis-29036 138 40 . . PUNCT fcis-29036 139 1 inadequate inadequate ADJ fcis-29036 139 2 impurity impurity NOUN fcis-29036 139 3 detection detection NOUN fcis-29036 139 4 not not PART fcis-29036 139 5 only only ADV fcis-29036 139 6 affects affect VERB fcis-29036 139 7 product product NOUN fcis-29036 139 8 quality quality NOUN fcis-29036 139 9 but but CCONJ fcis-29036 139 10 also also ADV fcis-29036 139 11 poses pose VERB fcis-29036 139 12 potential potential ADJ fcis-29036 139 13 food food NOUN fcis-29036 139 14 safety safety NOUN fcis-29036 139 15 risks risk NOUN fcis-29036 139 16 . . PUNCT fcis-29036 140 1 to to PART fcis-29036 140 2 address address VERB fcis-29036 140 3 this this PRON fcis-29036 140 4 , , PUNCT fcis-29036 140 5 this this DET fcis-29036 140 6 study study NOUN fcis-29036 140 7 designs design NOUN fcis-29036 140 8 and and CCONJ fcis-29036 140 9 proposes propose VERB fcis-29036 140 10 an an DET fcis-29036 140 11 impurity impurity NOUN fcis-29036 140 12 detection detection NOUN fcis-29036 140 13 system system NOUN fcis-29036 140 14 for for ADP fcis-29036 140 15 white white ADJ fcis-29036 140 16 sugar sugar NOUN fcis-29036 140 17 that that PRON fcis-29036 140 18 combines combine VERB fcis-29036 140 19 multimodal multimodal NOUN fcis-29036 140 20 image image NOUN fcis-29036 140 21 fusion fusion NOUN fcis-29036 140 22 technology technology NOUN fcis-29036 140 23 , , PUNCT fcis-29036 140 24 integrating integrate VERB fcis-29036 140 25 both both CCONJ fcis-29036 140 26 multispectral multispectral ADJ fcis-29036 140 27 and and CCONJ fcis-29036 140 28 rgb rgb PROPN fcis-29036 140 29 cameras camera NOUN fcis-29036 140 30 , , PUNCT fcis-29036 140 31 along along ADP fcis-29036 140 32 with with ADP fcis-29036 140 33 deep deep ADJ fcis-29036 140 34 learning learning NOUN fcis-29036 140 35 models model NOUN fcis-29036 140 36 , , PUNCT fcis-29036 140 37 for for ADP fcis-29036 140 38 automated automate VERB fcis-29036 140 39 , , PUNCT fcis-29036 140 40 non non ADJ fcis-29036 140 41 - - ADJ fcis-29036 140 42 destructive destructive ADJ fcis-29036 140 43 impurity impurity NOUN fcis-29036 140 44 detection detection NOUN fcis-29036 140 45 . . PUNCT fcis-29036 141 1 to to PART fcis-29036 141 2 better well ADV fcis-29036 141 3 utilize utilize VERB fcis-29036 141 4 multispectral multispectral ADJ fcis-29036 141 5 data datum NOUN fcis-29036 141 6 , , PUNCT fcis-29036 141 7 the the DET fcis-29036 141 8 traditional traditional ADJ fcis-29036 141 9 attention attention NOUN fcis-29036 141 10 unet unet NOUN fcis-29036 141 11 model model NOUN fcis-29036 141 12 was be AUX fcis-29036 141 13 optimized optimize VERB fcis-29036 141 14 by by ADP fcis-29036 141 15 expanding expand VERB fcis-29036 141 16 the the DET fcis-29036 141 17 input input NOUN fcis-29036 141 18 channels channel NOUN fcis-29036 141 19 from from ADP fcis-29036 141 20 the the DET fcis-29036 141 21 conventional conventional ADJ fcis-29036 141 22 single single ADJ fcis-29036 141 23 - - PUNCT fcis-29036 141 24 channel channel NOUN fcis-29036 141 25 grayscale grayscale NOUN fcis-29036 141 26 or or CCONJ fcis-29036 141 27 threechannel threechannel NOUN fcis-29036 141 28 rgb rgb PROPN fcis-29036 141 29 images image NOUN fcis-29036 141 30 to to ADP fcis-29036 141 31 thirteen thirteen NUM fcis-29036 141 32 channels channel NOUN fcis-29036 141 33 , , PUNCT fcis-29036 141 34 which which PRON fcis-29036 141 35 are be AUX fcis-29036 141 36 adapted adapt VERB fcis-29036 141 37 to to ADP fcis-29036 141 38 the the DET fcis-29036 141 39 system system NOUN fcis-29036 141 40 's 's PART fcis-29036 141 41 needs need NOUN fcis-29036 141 42 . . PUNCT fcis-29036 142 1 the the DET fcis-29036 142 2 multispectral multispectral ADJ fcis-29036 142 3 images image NOUN fcis-29036 142 4 cover cover VERB fcis-29036 142 5 multiple multiple ADJ fcis-29036 142 6 bands band NOUN fcis-29036 142 7 from from ADP fcis-29036 142 8 visible visible ADJ fcis-29036 142 9 light light NOUN fcis-29036 142 10 to to ADP fcis-29036 142 11 near near ADV fcis-29036 142 12 - - PUNCT fcis-29036 142 13 infrared infrare VERB fcis-29036 142 14 , , PUNCT fcis-29036 142 15 enabling enable VERB fcis-29036 142 16 the the DET fcis-29036 142 17 capture capture NOUN fcis-29036 142 18 of of ADP fcis-29036 142 19 both both CCONJ fcis-29036 142 20 chemical chemical ADJ fcis-29036 142 21 and and CCONJ fcis-29036 142 22 physical physical ADJ fcis-29036 142 23 characteristics characteristic NOUN fcis-29036 142 24 of of ADP fcis-29036 142 25 sugar sugar NOUN fcis-29036 142 26 and and CCONJ fcis-29036 142 27 its its PRON fcis-29036 142 28 impurities impurity NOUN fcis-29036 142 29 , , PUNCT fcis-29036 142 30 thereby thereby ADV fcis-29036 142 31 significantly significantly ADV fcis-29036 142 32 improving improve VERB fcis-29036 142 33 impurity impurity NOUN fcis-29036 142 34 detection detection NOUN fcis-29036 142 35 accuracy accuracy NOUN fcis-29036 142 36 . . PUNCT fcis-29036 143 1 experimental experimental ADJ fcis-29036 143 2 results result NOUN fcis-29036 143 3 demonstrate demonstrate VERB fcis-29036 143 4 a a DET fcis-29036 143 5 significant significant ADJ fcis-29036 143 6 performance performance NOUN fcis-29036 143 7 improvement improvement NOUN fcis-29036 143 8 in in ADP fcis-29036 143 9 the the DET fcis-29036 143 10 recognition recognition NOUN fcis-29036 143 11 of of ADP fcis-29036 143 12 impurity impurity NOUN fcis-29036 143 13 categories category NOUN fcis-29036 143 14 with with ADP fcis-29036 143 15 the the DET fcis-29036 143 16 optimized optimize VERB fcis-29036 143 17 model model NOUN fcis-29036 143 18 . . PUNCT fcis-29036 144 1 this this PRON fcis-29036 144 2 indicates indicate VERB fcis-29036 144 3 that that SCONJ fcis-29036 144 4 multispectral multispectral ADJ fcis-29036 144 5 data datum NOUN fcis-29036 144 6 helps help VERB fcis-29036 144 7 enhance enhance VERB fcis-29036 144 8 the the DET fcis-29036 144 9 precision precision NOUN fcis-29036 144 10 and and CCONJ fcis-29036 144 11 reliability reliability NOUN fcis-29036 144 12 of of ADP fcis-29036 144 13 impurity impurity NOUN fcis-29036 144 14 detection detection NOUN fcis-29036 144 15 . . PUNCT fcis-29036 145 1 the the DET fcis-29036 145 2 system system NOUN fcis-29036 145 3 provides provide VERB fcis-29036 145 4 a a DET fcis-29036 145 5 new new ADJ fcis-29036 145 6 solution solution NOUN fcis-29036 145 7 for for ADP fcis-29036 145 8 impurity impurity NOUN fcis-29036 145 9 detection detection NOUN fcis-29036 145 10 in in ADP fcis-29036 145 11 sugar sugar NOUN fcis-29036 145 12 production production NOUN fcis-29036 145 13 , , PUNCT fcis-29036 145 14 greatly greatly ADV fcis-29036 145 15 improving improve VERB fcis-29036 145 16 detection detection NOUN fcis-29036 145 17 efficiency efficiency NOUN fcis-29036 145 18 , , PUNCT fcis-29036 145 19 reducing reduce VERB fcis-29036 145 20 reliance reliance NOUN fcis-29036 145 21 on on ADP fcis-29036 145 22 manual manual ADJ fcis-29036 145 23 intervention intervention NOUN fcis-29036 145 24 , , PUNCT fcis-29036 145 25 and and CCONJ fcis-29036 145 26 enabling enable VERB fcis-29036 145 27 efficient efficient ADJ fcis-29036 145 28 real real ADJ fcis-29036 145 29 - - PUNCT fcis-29036 145 30 time time NOUN fcis-29036 145 31 monitoring monitoring NOUN fcis-29036 145 32 . . PUNCT fcis-29036 146 1 the the DET fcis-29036 146 2 system system NOUN fcis-29036 146 3 can can AUX fcis-29036 146 4 substantially substantially ADV fcis-29036 146 5 reduce reduce VERB fcis-29036 146 6 errors error NOUN fcis-29036 146 7 caused cause VERB fcis-29036 146 8 by by ADP fcis-29036 146 9 human human ADJ fcis-29036 146 10 factors factor NOUN fcis-29036 146 11 in in ADP fcis-29036 146 12 production production NOUN fcis-29036 146 13 , , PUNCT fcis-29036 146 14 lowering lower VERB fcis-29036 146 15 labor labor NOUN fcis-29036 146 16 costs cost NOUN fcis-29036 146 17 while while SCONJ fcis-29036 146 18 ensuring ensure VERB fcis-29036 146 19 food food NOUN fcis-29036 146 20 safety safety NOUN fcis-29036 146 21 during during ADP fcis-29036 146 22 the the DET fcis-29036 146 23 production production NOUN fcis-29036 146 24 process process NOUN fcis-29036 146 25 . . PUNCT fcis-29036 147 1 future future ADJ fcis-29036 147 2 research research NOUN fcis-29036 147 3 could could AUX fcis-29036 147 4 explore explore VERB fcis-29036 147 5 detection detection NOUN fcis-29036 147 6 methods method NOUN fcis-29036 147 7 for for ADP fcis-29036 147 8 different different ADJ fcis-29036 147 9 types type NOUN fcis-29036 147 10 of of ADP fcis-29036 147 11 impurities impurity NOUN fcis-29036 147 12 and and CCONJ fcis-29036 147 13 attempt attempt VERB fcis-29036 147 14 to to PART fcis-29036 147 15 apply apply VERB fcis-29036 147 16 the the DET fcis-29036 147 17 system system NOUN fcis-29036 147 18 to to ADP fcis-29036 147 19 other other ADJ fcis-29036 147 20 food food NOUN fcis-29036 147 21 industries industry NOUN fcis-29036 147 22 , , PUNCT fcis-29036 147 23 expanding expand VERB fcis-29036 147 24 its its PRON fcis-29036 147 25 application application NOUN fcis-29036 147 26 range range NOUN fcis-29036 147 27 . . PUNCT fcis-29036 148 1 acknowledgments acknowledgment NOUN fcis-29036 148 2 national national PROPN fcis-29036 148 3 natural natural PROPN fcis-29036 148 4 science science PROPN fcis-29036 148 5 foundation foundation PROPN fcis-29036 148 6 of of ADP fcis-29036 148 7 china china PROPN fcis-29036 148 8 ( ( PUNCT fcis-29036 148 9 11574058 11574058 NUM fcis-29036 148 10 ) ) PUNCT fcis-29036 148 11 . . PUNCT fcis-29036 149 1 references reference NOUN fcis-29036 149 2 [ [ X fcis-29036 149 3 1 1 NUM fcis-29036 149 4 ] ] X fcis-29036 149 5 müller müller PROPN fcis-29036 149 6 , , PUNCT fcis-29036 149 7 h. h. PROPN fcis-29036 149 8 , , PUNCT fcis-29036 149 9 & & CCONJ fcis-29036 149 10 agustín agustín PROPN fcis-29036 149 11 , , PUNCT fcis-29036 149 12 p. p. NOUN fcis-29036 149 13 ( ( PUNCT fcis-29036 149 14 2021 2021 NUM fcis-29036 149 15 ) ) PUNCT fcis-29036 149 16 . . PUNCT fcis-29036 150 1 a a DET fcis-29036 150 2 review review NOUN fcis-29036 150 3 of of ADP fcis-29036 150 4 traditional traditional ADJ fcis-29036 150 5 food food NOUN fcis-29036 150 6 quality quality NOUN fcis-29036 150 7 control control NOUN fcis-29036 150 8 methods method NOUN fcis-29036 150 9 and and CCONJ fcis-29036 150 10 their their PRON fcis-29036 150 11 limitations limitation NOUN fcis-29036 150 12 . . PUNCT fcis-29036 151 1 food food NOUN fcis-29036 151 2 chemistry chemistry NOUN fcis-29036 151 3 , , PUNCT fcis-29036 151 4 332 332 NUM fcis-29036 151 5 , , PUNCT fcis-29036 151 6 127373 127373 NUM fcis-29036 151 7 . . PUNCT fcis-29036 152 1 doi doi NOUN fcis-29036 152 2 : : PUNCT fcis-29036 152 3 10.1016 10.1016 NUM fcis-29036 152 4 / / SYM fcis-29036 152 5 j.foodchem.2020.127373 j.foodchem.2020.127373 NOUN fcis-29036 152 6 . . PUNCT fcis-29036 153 1 [ [ X fcis-29036 153 2 2 2 NUM fcis-29036 153 3 ] ] PUNCT fcis-29036 153 4 world world NOUN fcis-29036 153 5 health health NOUN fcis-29036 153 6 organization organization PROPN fcis-29036 153 7 . . PUNCT fcis-29036 154 1 guideline guideline NOUN fcis-29036 154 2 : : PUNCT fcis-29036 154 3 sugars sugar NOUN fcis-29036 154 4 intake intake NOUN fcis-29036 154 5 for for ADP fcis-29036 154 6 adults adult NOUN fcis-29036 154 7 and and CCONJ fcis-29036 154 8 children[m children[m NOUN fcis-29036 154 9 ] ] PUNCT fcis-29036 154 10 . . PUNCT fcis-29036 155 1 world world PROPN fcis-29036 155 2 health health PROPN fcis-29036 155 3 organization organization NOUN fcis-29036 155 4 , , PUNCT fcis-29036 155 5 2015 2015 NUM fcis-29036 155 6 . . PUNCT fcis-29036 156 1 [ [ X fcis-29036 156 2 3 3 X fcis-29036 156 3 ] ] PUNCT fcis-29036 156 4 jethi jethi PROPN fcis-29036 156 5 g g PROPN fcis-29036 156 6 s s PROPN fcis-29036 156 7 , , PUNCT fcis-29036 156 8 sunori sunori PROPN fcis-29036 156 9 s s PROPN fcis-29036 156 10 k k PROPN fcis-29036 156 11 , , PUNCT fcis-29036 156 12 joshi joshi PROPN fcis-29036 156 13 p p PROPN fcis-29036 156 14 , , PUNCT fcis-29036 156 15 et et PROPN fcis-29036 156 16 al al PROPN fcis-29036 156 17 . . PUNCT fcis-29036 156 18 application application NOUN fcis-29036 156 19 of of ADP fcis-29036 156 20 machine machine NOUN fcis-29036 156 21 learning learn VERB fcis-29036 156 22 in in ADP fcis-29036 156 23 assessing assess VERB fcis-29036 156 24 the the DET fcis-29036 156 25 sugar sugar NOUN fcis-29036 156 26 quality quality NOUN fcis-29036 156 27 in in ADP fcis-29036 156 28 sugar sugar NOUN fcis-29036 156 29 mills[c]// mills[c]// PROPN fcis-29036 156 30 2024 2024 NUM fcis-29036 156 31 5th 5th ADJ fcis-29036 156 32 international international ADJ fcis-29036 156 33 conference conference NOUN fcis-29036 156 34 on on ADP fcis-29036 156 35 intelligent intelligent ADJ fcis-29036 156 36 communication communication NOUN fcis-29036 156 37 technologies technology NOUN fcis-29036 156 38 and and CCONJ fcis-29036 156 39 virtual virtual ADJ fcis-29036 156 40 mobile mobile ADJ fcis-29036 156 41 networks network NOUN fcis-29036 156 42 ( ( PUNCT fcis-29036 156 43 icicv icicv PROPN fcis-29036 156 44 ) ) PUNCT fcis-29036 156 45 . . PUNCT fcis-29036 157 1 ieee ieee NOUN fcis-29036 157 2 , , PUNCT fcis-29036 157 3 2024 2024 NUM fcis-29036 157 4 : : PUNCT fcis-29036 157 5 167 167 NUM fcis-29036 157 6 - - SYM fcis-29036 157 7 172 172 NUM fcis-29036 157 8 . . PUNCT fcis-29036 158 1 [ [ X fcis-29036 158 2 4 4 NUM fcis-29036 158 3 ] ] X fcis-29036 158 4 wang wang PROPN fcis-29036 158 5 , , PUNCT fcis-29036 158 6 y. y. PROPN fcis-29036 158 7 h. h. PROPN fcis-29036 158 8 , , PUNCT fcis-29036 158 9 zhao zhao PROPN fcis-29036 158 10 , , PUNCT fcis-29036 158 11 d. d. PROPN fcis-29036 158 12 d. d. PROPN fcis-29036 158 13 , , PUNCT fcis-29036 158 14 & & CCONJ fcis-29036 158 15 li li PROPN fcis-29036 158 16 , , PUNCT fcis-29036 158 17 q. q. PROPN fcis-29036 158 18 t. t. PROPN fcis-29036 158 19 ( ( PUNCT fcis-29036 158 20 2014 2014 NUM fcis-29036 158 21 ) ) PUNCT fcis-29036 158 22 . . PUNCT fcis-29036 159 1 rapid rapid ADJ fcis-29036 159 2 analysis analysis NOUN fcis-29036 159 3 of of ADP fcis-29036 159 4 molasses molasse NOUN fcis-29036 159 5 hammer hammer NOUN fcis-29036 159 6 degree degree NOUN fcis-29036 159 7 , , PUNCT fcis-29036 159 8 sucrose sucrose NOUN fcis-29036 159 9 content content NOUN fcis-29036 159 10 , , PUNCT fcis-29036 159 11 and and CCONJ fcis-29036 159 12 reducing reduce VERB fcis-29036 159 13 sugar sugar NOUN fcis-29036 159 14 content content NOUN fcis-29036 159 15 by by ADP fcis-29036 159 16 near near ADV fcis-29036 159 17 - - PUNCT fcis-29036 159 18 infrared infrared ADJ fcis-29036 159 19 spectroscopy spectroscopy NOUN fcis-29036 159 20 [ [ X fcis-29036 159 21 in in ADP fcis-29036 159 22 chinese chinese PROPN fcis-29036 159 23 ] ] PUNCT fcis-29036 159 24 . . PUNCT fcis-29036 160 1 food food NOUN fcis-29036 160 2 science science NOUN fcis-29036 160 3 and and CCONJ fcis-29036 160 4 technology technology NOUN fcis-29036 160 5 , , PUNCT fcis-29036 160 6 39(9 39(9 NUM fcis-29036 160 7 ) ) PUNCT fcis-29036 160 8 , , PUNCT fcis-29036 160 9 284 284 NUM fcis-29036 160 10 - - SYM fcis-29036 160 11 288 288 NUM fcis-29036 160 12 . . PUNCT fcis-29036 161 1 [ [ X fcis-29036 161 2 5 5 X fcis-29036 161 3 ] ] X fcis-29036 161 4 hashim hashim PROPN fcis-29036 161 5 n n SYM fcis-29036 161 6 , , PUNCT fcis-29036 161 7 onwude onwude NOUN fcis-29036 161 8 d d PROPN fcis-29036 161 9 i i PROPN fcis-29036 161 10 , , PUNCT fcis-29036 161 11 osman osman PROPN fcis-29036 161 12 m m PROPN fcis-29036 161 13 s. s. PROPN fcis-29036 161 14 evaluation evaluation NOUN fcis-29036 161 15 of of ADP fcis-29036 161 16 chilling chilling ADJ fcis-29036 161 17 injury injury NOUN fcis-29036 161 18 in in ADP fcis-29036 161 19 mangoes mango NOUN fcis-29036 161 20 using use VERB fcis-29036 161 21 multispectral multispectral ADJ fcis-29036 161 22 imaging[j imaging[j NOUN fcis-29036 161 23 ] ] PUNCT fcis-29036 161 24 . . PUNCT fcis-29036 162 1 journal journal PROPN fcis-29036 162 2 of of ADP fcis-29036 162 3 food food NOUN fcis-29036 162 4 science science NOUN fcis-29036 162 5 , , PUNCT fcis-29036 162 6 2018 2018 NUM fcis-29036 162 7 , , PUNCT fcis-29036 162 8 83(5 83(5 NUM fcis-29036 162 9 ): ): PUNCT fcis-29036 162 10 1271 1271 NUM fcis-29036 162 11 - - SYM fcis-29036 162 12 1279 1279 NUM fcis-29036 162 13 . . PUNCT fcis-29036 163 1 [ [ X fcis-29036 163 2 6 6 NUM fcis-29036 163 3 ] ] X fcis-29036 163 4 huang huang PROPN fcis-29036 163 5 , , PUNCT fcis-29036 163 6 z. z. PROPN fcis-29036 163 7 , , PUNCT fcis-29036 163 8 chen chen PROPN fcis-29036 163 9 , , PUNCT fcis-29036 163 10 j. j. PROPN fcis-29036 163 11 , , PUNCT fcis-29036 163 12 tan tan PROPN fcis-29036 163 13 , , PUNCT fcis-29036 163 14 g. g. PROPN fcis-29036 163 15 , , PUNCT fcis-29036 163 16 et et PROPN fcis-29036 163 17 al al PROPN fcis-29036 163 18 . . PUNCT fcis-29036 163 19 ( ( PUNCT fcis-29036 163 20 2024 2024 NUM fcis-29036 163 21 ) ) PUNCT fcis-29036 163 22 . . PUNCT fcis-29036 163 23 hyperspectral hyperspectral ADJ fcis-29036 163 24 nontobacco nontobacco NOUN fcis-29036 163 25 material material NOUN fcis-29036 163 26 detection detection NOUN fcis-29036 163 27 method method NOUN fcis-29036 163 28 based base VERB fcis-29036 163 29 on on ADP fcis-29036 163 30 multi multi ADJ fcis-29036 163 31 - - ADJ fcis-29036 163 32 type type ADJ fcis-29036 163 33 convolutional convolutional ADJ fcis-29036 163 34 mixing mixing NOUN fcis-29036 163 35 [ [ X fcis-29036 163 36 in in ADP fcis-29036 163 37 chinese chinese PROPN fcis-29036 163 38 ] ] PUNCT fcis-29036 163 39 . . PUNCT fcis-29036 164 1 journal journal PROPN fcis-29036 164 2 of of ADP fcis-29036 164 3 image image NOUN fcis-29036 164 4 and and CCONJ fcis-29036 164 5 signal signal NOUN fcis-29036 164 6 processing processing NOUN fcis-29036 164 7 , , PUNCT fcis-29036 164 8 13 13 NUM fcis-29036 164 9 , , PUNCT fcis-29036 164 10 76 76 NUM fcis-29036 164 11 . . PUNCT fcis-29036 165 1 [ [ X fcis-29036 165 2 7 7 NUM fcis-29036 165 3 ] ] X fcis-29036 165 4 russell russell PROPN fcis-29036 165 5 , , PUNCT fcis-29036 165 6 b. b. PROPN fcis-29036 165 7 c. c. PROPN fcis-29036 165 8 , , PUNCT fcis-29036 165 9 torralba torralba PROPN fcis-29036 165 10 , , PUNCT fcis-29036 165 11 a. a. NOUN fcis-29036 165 12 , , PUNCT fcis-29036 165 13 murphy murphy PROPN fcis-29036 165 14 , , PUNCT fcis-29036 165 15 k. k. PROPN fcis-29036 165 16 p. p. PROPN fcis-29036 165 17 , , PUNCT fcis-29036 165 18 & & CCONJ fcis-29036 165 19 freeman freeman PROPN fcis-29036 165 20 , , PUNCT fcis-29036 165 21 w. w. PROPN fcis-29036 165 22 t. t. PROPN fcis-29036 165 23 ( ( PUNCT fcis-29036 165 24 200˜ 200˜ NUM fcis-29036 165 25 ) ) PUNCT fcis-29036 165 26 . . PUNCT fcis-29036 166 1 labelme labelme PROPN fcis-29036 166 2 : : PUNCT fcis-29036 166 3 a a DET fcis-29036 166 4 database database NOUN fcis-29036 166 5 and and CCONJ fcis-29036 166 6 web web NOUN fcis-29036 166 7 - - PUNCT fcis-29036 166 8 based base VERB fcis-29036 166 9 tool tool NOUN fcis-29036 166 10 for for ADP fcis-29036 166 11 image image NOUN fcis-29036 166 12 annotation annotation NOUN fcis-29036 166 13 . . PUNCT fcis-29036 167 1 international international ADJ fcis-29036 167 2 journal journal NOUN fcis-29036 167 3 of of ADP fcis-29036 167 4 computer computer NOUN fcis-29036 167 5 vision vision NOUN fcis-29036 167 6 , , PUNCT fcis-29036 167 7 77(1 77(1 NOUN fcis-29036 167 8 - - SYM fcis-29036 167 9 3 3 NUM fcis-29036 167 10 ) ) PUNCT fcis-29036 167 11 , , PUNCT fcis-29036 167 12 157 157 NUM fcis-29036 167 13 - - SYM fcis-29036 167 14 173 173 NUM fcis-29036 167 15 . . PUNCT fcis-29036 168 1 [ [ X fcis-29036 168 2 8 8 NUM fcis-29036 168 3 ] ] SYM fcis-29036 168 4 ronneberger ronneberger NOUN fcis-29036 168 5 , , PUNCT fcis-29036 168 6 o. o. PROPN fcis-29036 168 7 , , PUNCT fcis-29036 168 8 fischer fischer PROPN fcis-29036 168 9 , , PUNCT fcis-29036 168 10 p. p. PROPN fcis-29036 168 11 , , PUNCT fcis-29036 168 12 & & CCONJ fcis-29036 168 13 brox brox PROPN fcis-29036 168 14 , , PUNCT fcis-29036 168 15 t. t. PROPN fcis-29036 168 16 ( ( PUNCT fcis-29036 168 17 2015 2015 NUM fcis-29036 168 18 ) ) PUNCT fcis-29036 168 19 . . PUNCT fcis-29036 169 1 u u NOUN fcis-29036 169 2 - - NOUN fcis-29036 169 3 net net ADJ fcis-29036 169 4 : : PUNCT fcis-29036 169 5 convolutional convolutional ADJ fcis-29036 169 6 networks network NOUN fcis-29036 169 7 for for ADP fcis-29036 169 8 biomedical biomedical ADJ fcis-29036 169 9 image image NOUN fcis-29036 169 10 segmentation segmentation NOUN fcis-29036 169 11 . . PUNCT fcis-29036 170 1 medical medical ADJ fcis-29036 170 2 image image NOUN fcis-29036 170 3 computing computing NOUN fcis-29036 170 4 and and CCONJ fcis-29036 170 5 computer computer NOUN fcis-29036 170 6 - - PUNCT fcis-29036 170 7 assisted assist VERB fcis-29036 170 8 intervention intervention NOUN fcis-29036 170 9 ( ( PUNCT fcis-29036 170 10 miccai miccai PROPN fcis-29036 170 11 ) ) PUNCT fcis-29036 170 12 , , PUNCT fcis-29036 170 13 pp pp PROPN fcis-29036 170 14 . . PUNCT fcis-29036 171 1 234 234 NUM fcis-29036 171 2 - - SYM fcis-29036 171 3 241 241 NUM fcis-29036 171 4 . . PUNCT fcis-29036 171 5 springer springer PROPN fcis-29036 171 6 , , PUNCT fcis-29036 171 7 cham cham PROPN fcis-29036 171 8 . . PUNCT fcis-29036 172 1 doi doi NOUN fcis-29036 172 2 : : PUNCT fcis-29036 172 3 10 10 NUM fcis-29036 172 4 . . X fcis-29036 173 1 1007/ 1007/ NUM fcis-29036 173 2 978 978 NUM fcis-29036 173 3 - - SYM fcis-29036 173 4 3 3 NUM fcis-29036 173 5 - - NUM fcis-29036 173 6 319 319 NUM fcis-29036 173 7 - - PUNCT fcis-29036 173 8 24574 24574 NUM fcis-29036 173 9 - - SYM fcis-29036 173 10 4_28 4_28 NOUN fcis-29036 173 11 . . PUNCT fcis-29036 174 1 [ [ X fcis-29036 174 2 9 9 NUM fcis-29036 174 3 ] ] SYM fcis-29036 174 4 oktay oktay PROPN fcis-29036 174 5 , , PUNCT fcis-29036 174 6 o. o. PROPN fcis-29036 174 7 , , PUNCT fcis-29036 174 8 schlemper schlemper ADJ fcis-29036 174 9 , , PUNCT fcis-29036 174 10 j. j. PROPN fcis-29036 174 11 , , PUNCT fcis-29036 174 12 folgoc folgoc PROPN fcis-29036 174 13 , , PUNCT fcis-29036 174 14 l. l. PROPN fcis-29036 174 15 l. l. PROPN fcis-29036 174 16 , , PUNCT fcis-29036 174 17 et et PROPN fcis-29036 174 18 al al PROPN fcis-29036 174 19 . . PUNCT fcis-29036 175 1 ( ( PUNCT fcis-29036 175 2 2018 2018 NUM fcis-29036 175 3 ) ) PUNCT fcis-29036 175 4 . . PUNCT fcis-29036 176 1 attention attention NOUN fcis-29036 176 2 u u NOUN fcis-29036 176 3 - - NOUN fcis-29036 176 4 net net ADJ fcis-29036 176 5 : : PUNCT fcis-29036 176 6 learning learn VERB fcis-29036 176 7 where where SCONJ fcis-29036 176 8 to to PART fcis-29036 176 9 look look VERB fcis-29036 176 10 for for ADP fcis-29036 176 11 the the DET fcis-29036 176 12 pancreas pancrea NOUN fcis-29036 176 13 . . PUNCT fcis-29036 177 1 arxiv arxiv PROPN fcis-29036 177 2 preprint preprint NOUN fcis-29036 177 3 arxiv:1804.03999 arxiv:1804.03999 NOUN fcis-29036 177 4 . . PUNCT fcis-29036 178 1 [ [ X fcis-29036 178 2 10 10 NUM fcis-29036 178 3 ] ] X fcis-29036 178 4 wang wang PROPN fcis-29036 178 5 , , PUNCT fcis-29036 178 6 x. x. PROPN fcis-29036 178 7 , , PUNCT fcis-29036 178 8 & & CCONJ fcis-29036 178 9 guo guo PROPN fcis-29036 178 10 , , PUNCT fcis-29036 178 11 x. x. NOUN fcis-29036 178 12 ( ( PUNCT fcis-29036 178 13 2020 2020 NUM fcis-29036 178 14 ) ) PUNCT fcis-29036 178 15 . . PUNCT fcis-29036 179 1 attention attention NOUN fcis-29036 179 2 u u NOUN fcis-29036 179 3 - - NOUN fcis-29036 179 4 net net NOUN fcis-29036 179 5 based base VERB fcis-29036 179 6 on on ADP fcis-29036 179 7 residual residual ADJ fcis-29036 179 8 learning learning NOUN fcis-29036 179 9 for for ADP fcis-29036 179 10 medical medical ADJ fcis-29036 179 11 image image NOUN fcis-29036 179 12 segmentation segmentation NOUN fcis-29036 179 13 . . PUNCT fcis-29036 180 1 ieee ieee NOUN fcis-29036 180 2 access access NOUN fcis-29036 180 3 , , PUNCT fcis-29036 180 4 8 8 NUM fcis-29036 180 5 , , PUNCT fcis-29036 180 6 55230 55230 NUM fcis-29036 180 7 - - SYM fcis-29036 180 8 55239 55239 NUM fcis-29036 180 9 . . PUNCT fcis-29036 181 1 doi doi NOUN fcis-29036 181 2 : : PUNCT fcis-29036 181 3 10.1109/ 10.1109/ NUM fcis-29036 181 4 access access NOUN fcis-29036 181 5 . . PUNCT fcis-29036 182 1 2020 2020 NUM fcis-29036 182 2 . . PUNCT fcis-29036 183 1 298 298 NUM fcis-29036 183 2 2408 2408 NUM fcis-29036 183 3 . . PUNCT