id sid tid token lemma pos fcis-12446 1 1 frontiers frontier NOUN fcis-12446 1 2 in in ADP fcis-12446 1 3 computing computing NOUN fcis-12446 1 4 and and CCONJ fcis-12446 1 5 intelligent intelligent ADJ fcis-12446 1 6 systems system NOUN fcis-12446 1 7 issn issn VERB fcis-12446 1 8 : : PUNCT fcis-12446 1 9 2832 2832 NUM fcis-12446 1 10 - - SYM fcis-12446 1 11 6024 6024 NUM fcis-12446 1 12 | | NOUN fcis-12446 1 13 vol vol NOUN fcis-12446 1 14 . . PROPN fcis-12446 2 1 5 5 NUM fcis-12446 2 2 , , PUNCT fcis-12446 2 3 no no INTJ fcis-12446 2 4 . . NOUN fcis-12446 2 5 2 2 NUM fcis-12446 2 6 , , PUNCT fcis-12446 2 7 2023 2023 NUM fcis-12446 2 8 61 61 NUM fcis-12446 2 9 aphid aphid NOUN fcis-12446 2 10 detection detection NOUN fcis-12446 2 11 network network NOUN fcis-12446 2 12 from from ADP fcis-12446 2 13 global global ADJ fcis-12446 2 14 to to ADP fcis-12446 2 15 local local ADJ fcis-12446 2 16 hui hui PROPN fcis-12446 2 17 zhang zhang PROPN fcis-12446 2 18 1 1 NUM fcis-12446 2 19 , , PUNCT fcis-12446 2 20 2 2 NUM fcis-12446 2 21 , , PUNCT fcis-12446 2 22 * * PUNCT fcis-12446 2 23 , , PUNCT fcis-12446 2 24 xiaoping xiaoping PROPN fcis-12446 2 25 yang yang PROPN fcis-12446 2 26 1 1 NUM fcis-12446 2 27 1 1 NUM fcis-12446 2 28 zhejiang zhejiang PROPN fcis-12446 2 29 yuexiu yuexiu PROPN fcis-12446 2 30 university university PROPN fcis-12446 2 31 , , PUNCT fcis-12446 2 32 shaoxing shaoxing PROPN fcis-12446 2 33 zhejiang zhejiang PROPN fcis-12446 2 34 , , PUNCT fcis-12446 2 35 312000 312000 NUM fcis-12446 2 36 , , PUNCT fcis-12446 2 37 china china PROPN fcis-12446 2 38 2 2 NUM fcis-12446 2 39 school school NOUN fcis-12446 2 40 of of ADP fcis-12446 2 41 science science NOUN fcis-12446 2 42 , , PUNCT fcis-12446 2 43 zhejiang zhejiang PROPN fcis-12446 2 44 university university PROPN fcis-12446 2 45 of of ADP fcis-12446 2 46 science science NOUN fcis-12446 2 47 and and CCONJ fcis-12446 2 48 technology technology NOUN fcis-12446 2 49 , , PUNCT fcis-12446 2 50 hangzhou hangzhou PROPN fcis-12446 2 51 zhejiang zhejiang PROPN fcis-12446 2 52 , , PUNCT fcis-12446 2 53 310023 310023 NUM fcis-12446 2 54 , , PUNCT fcis-12446 2 55 china china PROPN fcis-12446 2 56 * * PUNCT fcis-12446 2 57 corresponding correspond VERB fcis-12446 2 58 author author NOUN fcis-12446 2 59 : : PUNCT fcis-12446 2 60 hui hui PROPN fcis-12446 2 61 zhang zhang PROPN fcis-12446 2 62 ( ( PUNCT fcis-12446 2 63 email email NOUN fcis-12446 2 64 : : PUNCT fcis-12446 2 65 1520939382@qq.com 1520939382@qq.com NUM fcis-12446 2 66 ) ) PUNCT fcis-12446 2 67 abstract abstract NOUN fcis-12446 2 68 : : PUNCT fcis-12446 2 69 common common ADJ fcis-12446 2 70 aphids aphid NOUN fcis-12446 2 71 on on ADP fcis-12446 2 72 crops crop NOUN fcis-12446 2 73 are be AUX fcis-12446 2 74 not not PART fcis-12446 2 75 suitable suitable ADJ fcis-12446 2 76 for for ADP fcis-12446 2 77 general general ADJ fcis-12446 2 78 - - PUNCT fcis-12446 2 79 purpose purpose NOUN fcis-12446 2 80 object object NOUN fcis-12446 2 81 detection detection NOUN fcis-12446 2 82 frameworks framework NOUN fcis-12446 2 83 due due ADP fcis-12446 2 84 to to ADP fcis-12446 2 85 their their PRON fcis-12446 2 86 small small ADJ fcis-12446 2 87 size size NOUN fcis-12446 2 88 and and CCONJ fcis-12446 2 89 the the DET fcis-12446 2 90 presence presence NOUN fcis-12446 2 91 of of ADP fcis-12446 2 92 occlusion occlusion NOUN fcis-12446 2 93 . . PUNCT fcis-12446 3 1 taking take VERB fcis-12446 3 2 this this PRON fcis-12446 3 3 as as ADP fcis-12446 3 4 a a DET fcis-12446 3 5 starting starting NOUN fcis-12446 3 6 point point NOUN fcis-12446 3 7 , , PUNCT fcis-12446 3 8 we we PRON fcis-12446 3 9 fully fully ADV fcis-12446 3 10 consider consider VERB fcis-12446 3 11 the the DET fcis-12446 3 12 characteristics characteristic NOUN fcis-12446 3 13 of of ADP fcis-12446 3 14 aphid aphid NOUN fcis-12446 3 15 targets target NOUN fcis-12446 3 16 and and CCONJ fcis-12446 3 17 propose propose VERB fcis-12446 3 18 a a DET fcis-12446 3 19 network network NOUN fcis-12446 3 20 called call VERB fcis-12446 3 21 overall overall ADJ fcis-12446 3 22 - - PUNCT fcis-12446 3 23 specific specific ADJ fcis-12446 3 24 net net NOUN fcis-12446 3 25 ( ( PUNCT fcis-12446 3 26 os os NOUN fcis-12446 3 27 - - PUNCT fcis-12446 3 28 net net NOUN fcis-12446 3 29 ) ) PUNCT fcis-12446 3 30 for for ADP fcis-12446 3 31 detecting detect VERB fcis-12446 3 32 aphid aphid NOUN fcis-12446 3 33 targets target NOUN fcis-12446 3 34 in in ADP fcis-12446 3 35 a a DET fcis-12446 3 36 manner manner NOUN fcis-12446 3 37 that that PRON fcis-12446 3 38 goes go VERB fcis-12446 3 39 from from ADP fcis-12446 3 40 overall overall ADV fcis-12446 3 41 to to ADP fcis-12446 3 42 specific specific ADJ fcis-12446 3 43 in in ADP fcis-12446 3 44 a a DET fcis-12446 3 45 single single ADJ fcis-12446 3 46 - - PUNCT fcis-12446 3 47 stage stage NOUN fcis-12446 3 48 detection detection NOUN fcis-12446 3 49 network network NOUN fcis-12446 3 50 . . PUNCT fcis-12446 4 1 this this DET fcis-12446 4 2 network network NOUN fcis-12446 4 3 consists consist VERB fcis-12446 4 4 of of ADP fcis-12446 4 5 two two NUM fcis-12446 4 6 detection detection NOUN fcis-12446 4 7 head head NOUN fcis-12446 4 8 modules module NOUN fcis-12446 4 9 . . PUNCT fcis-12446 5 1 the the DET fcis-12446 5 2 first first ADJ fcis-12446 5 3 detection detection NOUN fcis-12446 5 4 head head NOUN fcis-12446 5 5 module module NOUN fcis-12446 5 6 is be AUX fcis-12446 5 7 responsible responsible ADJ fcis-12446 5 8 for for ADP fcis-12446 5 9 detecting detect VERB fcis-12446 5 10 aphids aphid NOUN fcis-12446 5 11 in in ADP fcis-12446 5 12 densely densely ADV fcis-12446 5 13 populated populate VERB fcis-12446 5 14 areas area NOUN fcis-12446 5 15 , , PUNCT fcis-12446 5 16 while while SCONJ fcis-12446 5 17 the the DET fcis-12446 5 18 second second ADJ fcis-12446 5 19 module module NOUN fcis-12446 5 20 , , PUNCT fcis-12446 5 21 based base VERB fcis-12446 5 22 on on ADP fcis-12446 5 23 the the DET fcis-12446 5 24 first first ADJ fcis-12446 5 25 module module NOUN fcis-12446 5 26 , , PUNCT fcis-12446 5 27 detects detect NOUN fcis-12446 5 28 aphids aphid NOUN fcis-12446 5 29 in in ADP fcis-12446 5 30 densely densely ADV fcis-12446 5 31 populated populate VERB fcis-12446 5 32 areas area NOUN fcis-12446 5 33 by by ADP fcis-12446 5 34 deploying deploy VERB fcis-12446 5 35 denser dense ADJ fcis-12446 5 36 anchor anchor NOUN fcis-12446 5 37 boxes box NOUN fcis-12446 5 38 . . PUNCT fcis-12446 6 1 experimental experimental ADJ fcis-12446 6 2 results result NOUN fcis-12446 6 3 on on ADP fcis-12446 6 4 our our PRON fcis-12446 6 5 dataset dataset NOUN fcis-12446 6 6 show show NOUN fcis-12446 6 7 that that SCONJ fcis-12446 6 8 the the DET fcis-12446 6 9 average average ADJ fcis-12446 6 10 accuracy accuracy NOUN fcis-12446 6 11 can can AUX fcis-12446 6 12 be be AUX fcis-12446 6 13 improved improve VERB fcis-12446 6 14 by by ADP fcis-12446 6 15 approximately approximately ADV fcis-12446 6 16 5.1 5.1 NUM fcis-12446 6 17 % % NOUN fcis-12446 6 18 compared compare VERB fcis-12446 6 19 to to ADP fcis-12446 6 20 the the DET fcis-12446 6 21 baseline baseline NOUN fcis-12446 6 22 network network NOUN fcis-12446 6 23 . . PUNCT fcis-12446 7 1 keywords keyword NOUN fcis-12446 7 2 : : PUNCT fcis-12446 7 3 deep deep ADJ fcis-12446 7 4 learning learning NOUN fcis-12446 7 5 ; ; PUNCT fcis-12446 7 6 convolutional convolutional ADJ fcis-12446 7 7 neural neural ADJ fcis-12446 7 8 networks network NOUN fcis-12446 7 9 ; ; PUNCT fcis-12446 7 10 pest pest VERB fcis-12446 7 11 detection detection NOUN fcis-12446 7 12 . . PUNCT fcis-12446 8 1 1 1 X fcis-12446 8 2 . . X fcis-12446 8 3 introduction introduction NOUN fcis-12446 8 4 agriculture agriculture NOUN fcis-12446 8 5 is be AUX fcis-12446 8 6 the the DET fcis-12446 8 7 foundation foundation NOUN fcis-12446 8 8 of of ADP fcis-12446 8 9 the the DET fcis-12446 8 10 national national ADJ fcis-12446 8 11 economy economy NOUN fcis-12446 8 12 , , PUNCT fcis-12446 8 13 and and CCONJ fcis-12446 8 14 crop crop NOUN fcis-12446 8 15 production production NOUN fcis-12446 8 16 is be AUX fcis-12446 8 17 crucial crucial ADJ fcis-12446 8 18 for for ADP fcis-12446 8 19 the the DET fcis-12446 8 20 livelihood livelihood NOUN fcis-12446 8 21 of of ADP fcis-12446 8 22 the the DET fcis-12446 8 23 people people NOUN fcis-12446 8 24 . . PUNCT fcis-12446 9 1 as as ADP fcis-12446 9 2 of of ADP fcis-12446 9 3 the the DET fcis-12446 9 4 end end NOUN fcis-12446 9 5 of of ADP fcis-12446 9 6 2022 2022 NUM fcis-12446 9 7 , , PUNCT fcis-12446 9 8 without without ADP fcis-12446 9 9 considering consider VERB fcis-12446 9 10 stockpiles stockpile NOUN fcis-12446 9 11 , , PUNCT fcis-12446 9 12 the the DET fcis-12446 9 13 per per ADP fcis-12446 9 14 capita capita X fcis-12446 9 15 grain grain NOUN fcis-12446 9 16 ownership ownership NOUN fcis-12446 9 17 of of ADP fcis-12446 9 18 china china PROPN fcis-12446 9 19 's 's PART fcis-12446 9 20 1.4 1.4 NUM fcis-12446 9 21 billion billion NUM fcis-12446 9 22 population population NOUN fcis-12446 9 23 reached reach VERB fcis-12446 9 24 475 475 NUM fcis-12446 9 25 kilograms kilogram NOUN fcis-12446 9 26 , , PUNCT fcis-12446 9 27 while while SCONJ fcis-12446 9 28 the the DET fcis-12446 9 29 internationally internationally ADV fcis-12446 9 30 recognized recognize VERB fcis-12446 9 31 food food NOUN fcis-12446 9 32 security security NOUN fcis-12446 9 33 line line NOUN fcis-12446 9 34 is be AUX fcis-12446 9 35 400 400 NUM fcis-12446 9 36 kilograms kilogram NOUN fcis-12446 9 37 per per ADP fcis-12446 9 38 capita capita NOUN fcis-12446 9 39 . . PUNCT fcis-12446 10 1 the the DET fcis-12446 10 2 average average ADJ fcis-12446 10 3 grain grain NOUN fcis-12446 10 4 ownership ownership NOUN fcis-12446 10 5 for for ADP fcis-12446 10 6 the the DET fcis-12446 10 7 global global ADJ fcis-12446 10 8 population population NOUN fcis-12446 10 9 of of ADP fcis-12446 10 10 8 8 NUM fcis-12446 10 11 billion billion NUM fcis-12446 10 12 is be AUX fcis-12446 10 13 351 351 NUM fcis-12446 10 14 kilograms kilogram NOUN fcis-12446 10 15 , , PUNCT fcis-12446 10 16 all all PRON fcis-12446 10 17 of of ADP fcis-12446 10 18 which which PRON fcis-12446 10 19 reflects reflect VERB fcis-12446 10 20 the the DET fcis-12446 10 21 significant significant ADJ fcis-12446 10 22 achievements achievement NOUN fcis-12446 10 23 in in ADP fcis-12446 10 24 the the DET fcis-12446 10 25 development development NOUN fcis-12446 10 26 of of ADP fcis-12446 10 27 china china PROPN fcis-12446 10 28 's 's PART fcis-12446 10 29 agricultural agricultural ADJ fcis-12446 10 30 industry industry NOUN fcis-12446 10 31 . . PUNCT fcis-12446 11 1 especially especially ADV fcis-12446 11 2 in in ADP fcis-12446 11 3 main main ADJ fcis-12446 11 4 grains grain NOUN fcis-12446 11 5 such such ADJ fcis-12446 11 6 as as ADP fcis-12446 11 7 rice rice NOUN fcis-12446 11 8 , , PUNCT fcis-12446 11 9 wheat wheat NOUN fcis-12446 11 10 , , PUNCT fcis-12446 11 11 and and CCONJ fcis-12446 11 12 corn corn NOUN fcis-12446 11 13 , , PUNCT fcis-12446 11 14 production production NOUN fcis-12446 11 15 and and CCONJ fcis-12446 11 16 demand demand NOUN fcis-12446 11 17 are be AUX fcis-12446 11 18 basically basically ADV fcis-12446 11 19 matched match VERB fcis-12446 11 20 . . PUNCT fcis-12446 12 1 although although SCONJ fcis-12446 12 2 the the DET fcis-12446 12 3 total total ADJ fcis-12446 12 4 grain grain NOUN fcis-12446 12 5 production production NOUN fcis-12446 12 6 increased increase VERB fcis-12446 12 7 by by ADP fcis-12446 12 8 0.5 0.5 NUM fcis-12446 12 9 % % NOUN fcis-12446 12 10 in in ADP fcis-12446 12 11 2022 2022 NUM fcis-12446 12 12 compared compare VERB fcis-12446 12 13 to to ADP fcis-12446 12 14 2021 2021 NUM fcis-12446 12 15 , , PUNCT fcis-12446 12 16 the the DET fcis-12446 12 17 average average ADJ fcis-12446 12 18 yield yield NOUN fcis-12446 12 19 per per ADP fcis-12446 12 20 acre acre NOUN fcis-12446 12 21 of of ADP fcis-12446 12 22 grain grain NOUN fcis-12446 12 23 decreased decrease VERB fcis-12446 12 24 by by ADP fcis-12446 12 25 1 1 NUM fcis-12446 12 26 % % NOUN fcis-12446 12 27 . . PUNCT fcis-12446 13 1 this this PRON fcis-12446 13 2 is be AUX fcis-12446 13 3 mainly mainly ADV fcis-12446 13 4 due due ADJ fcis-12446 13 5 to to ADP fcis-12446 13 6 factors factor NOUN fcis-12446 13 7 such such ADJ fcis-12446 13 8 as as ADP fcis-12446 13 9 weather weather NOUN fcis-12446 13 10 and and CCONJ fcis-12446 13 11 pests pest NOUN fcis-12446 13 12 . . PUNCT fcis-12446 14 1 however however ADV fcis-12446 14 2 , , PUNCT fcis-12446 14 3 the the DET fcis-12446 14 4 decrease decrease NOUN fcis-12446 14 5 in in ADP fcis-12446 14 6 production production NOUN fcis-12446 14 7 due due ADP fcis-12446 14 8 to to ADP fcis-12446 14 9 weather weather NOUN fcis-12446 14 10 reasons reason NOUN fcis-12446 14 11 is be AUX fcis-12446 14 12 uncontrollable uncontrollable ADJ fcis-12446 14 13 , , PUNCT fcis-12446 14 14 while while SCONJ fcis-12446 14 15 the the DET fcis-12446 14 16 prevention prevention NOUN fcis-12446 14 17 of of ADP fcis-12446 14 18 pests pest NOUN fcis-12446 14 19 can can AUX fcis-12446 14 20 be be AUX fcis-12446 14 21 planned plan VERB fcis-12446 14 22 . . PUNCT fcis-12446 15 1 pest pest NOUN fcis-12446 15 2 infestations infestation NOUN fcis-12446 15 3 occur occur VERB fcis-12446 15 4 throughout throughout ADP fcis-12446 15 5 the the DET fcis-12446 15 6 various various ADJ fcis-12446 15 7 stages stage NOUN fcis-12446 15 8 of of ADP fcis-12446 15 9 crop crop NOUN fcis-12446 15 10 growth growth NOUN fcis-12446 15 11 , , PUNCT fcis-12446 15 12 resulting result VERB fcis-12446 15 13 in in ADP fcis-12446 15 14 incalculable incalculable ADJ fcis-12446 15 15 grain grain NOUN fcis-12446 15 16 losses loss NOUN fcis-12446 15 17 . . PUNCT fcis-12446 16 1 therefore therefore ADV fcis-12446 16 2 , , PUNCT fcis-12446 16 3 how how SCONJ fcis-12446 16 4 to to PART fcis-12446 16 5 more more ADV fcis-12446 16 6 accurately accurately ADV fcis-12446 16 7 monitor monitor VERB fcis-12446 16 8 pest pest NOUN fcis-12446 16 9 targets target NOUN fcis-12446 16 10 and and CCONJ fcis-12446 16 11 apply apply VERB fcis-12446 16 12 pesticides pesticide NOUN fcis-12446 16 13 rationally rationally ADV fcis-12446 16 14 is be AUX fcis-12446 16 15 crucial crucial ADJ fcis-12446 16 16 to to ADP fcis-12446 16 17 protecting protect VERB fcis-12446 16 18 crop crop NOUN fcis-12446 16 19 growth growth NOUN fcis-12446 16 20 . . PUNCT fcis-12446 17 1 before before SCONJ fcis-12446 17 2 the the DET fcis-12446 17 3 widespread widespread ADJ fcis-12446 17 4 application application NOUN fcis-12446 17 5 of of ADP fcis-12446 17 6 computer computer NOUN fcis-12446 17 7 vision vision NOUN fcis-12446 17 8 , , PUNCT fcis-12446 17 9 crop crop NOUN fcis-12446 17 10 pest pest NOUN fcis-12446 17 11 monitoring monitoring NOUN fcis-12446 17 12 mainly mainly ADV fcis-12446 17 13 relied rely VERB fcis-12446 17 14 on on ADP fcis-12446 17 15 field field NOUN fcis-12446 17 16 observations observation NOUN fcis-12446 17 17 by by ADP fcis-12446 17 18 farmers farmer NOUN fcis-12446 17 19 and and CCONJ fcis-12446 17 20 agricultural agricultural ADJ fcis-12446 17 21 professionals professional NOUN fcis-12446 17 22 . . PUNCT fcis-12446 18 1 however however ADV fcis-12446 18 2 , , PUNCT fcis-12446 18 3 this this DET fcis-12446 18 4 monitoring monitoring NOUN fcis-12446 18 5 method method NOUN fcis-12446 18 6 has have VERB fcis-12446 18 7 two two NUM fcis-12446 18 8 drawbacks drawback NOUN fcis-12446 18 9 : : PUNCT fcis-12446 18 10 on on ADP fcis-12446 18 11 the the DET fcis-12446 18 12 one one NUM fcis-12446 18 13 hand hand NOUN fcis-12446 18 14 , , PUNCT fcis-12446 18 15 the the DET fcis-12446 18 16 observation observation NOUN fcis-12446 18 17 results result NOUN fcis-12446 18 18 are be AUX fcis-12446 18 19 subjective subjective ADJ fcis-12446 18 20 and and CCONJ fcis-12446 18 21 limited limited ADJ fcis-12446 18 22 , , PUNCT fcis-12446 18 23 requiring require VERB fcis-12446 18 24 a a DET fcis-12446 18 25 high high ADJ fcis-12446 18 26 level level NOUN fcis-12446 18 27 of of ADP fcis-12446 18 28 knowledge knowledge NOUN fcis-12446 18 29 from from ADP fcis-12446 18 30 relevant relevant ADJ fcis-12446 18 31 personnel personnel NOUN fcis-12446 18 32 ; ; PUNCT fcis-12446 18 33 on on ADP fcis-12446 18 34 the the DET fcis-12446 18 35 other other ADJ fcis-12446 18 36 hand hand NOUN fcis-12446 18 37 , , PUNCT fcis-12446 18 38 this this DET fcis-12446 18 39 process process NOUN fcis-12446 18 40 is be AUX fcis-12446 18 41 relatively relatively ADV fcis-12446 18 42 labor labor NOUN fcis-12446 18 43 - - PUNCT fcis-12446 18 44 intensive intensive ADJ fcis-12446 18 45 , , PUNCT fcis-12446 18 46 with with ADP fcis-12446 18 47 poor poor ADJ fcis-12446 18 48 working working NOUN fcis-12446 18 49 conditions condition NOUN fcis-12446 18 50 , , PUNCT fcis-12446 18 51 leading lead VERB fcis-12446 18 52 to to ADP fcis-12446 18 53 a a DET fcis-12446 18 54 reduction reduction NOUN fcis-12446 18 55 in in ADP fcis-12446 18 56 the the DET fcis-12446 18 57 number number NOUN fcis-12446 18 58 of of ADP fcis-12446 18 59 professionals professional NOUN fcis-12446 18 60 , , PUNCT fcis-12446 18 61 making make VERB fcis-12446 18 62 it it PRON fcis-12446 18 63 difficult difficult ADJ fcis-12446 18 64 to to PART fcis-12446 18 65 meet meet VERB fcis-12446 18 66 the the DET fcis-12446 18 67 needs need NOUN fcis-12446 18 68 of of ADP fcis-12446 18 69 largescale largescale NOUN fcis-12446 18 70 pest pest NOUN fcis-12446 18 71 monitoring monitoring NOUN fcis-12446 18 72 in in ADP fcis-12446 18 73 the the DET fcis-12446 18 74 country country NOUN fcis-12446 18 75 . . PUNCT fcis-12446 19 1 therefore therefore ADV fcis-12446 19 2 , , PUNCT fcis-12446 19 3 designing design VERB fcis-12446 19 4 a a DET fcis-12446 19 5 more more ADV fcis-12446 19 6 reasonable reasonable ADJ fcis-12446 19 7 and and CCONJ fcis-12446 19 8 efficient efficient ADJ fcis-12446 19 9 pest pest NOUN fcis-12446 19 10 monitoring monitoring NOUN fcis-12446 19 11 scheme scheme NOUN fcis-12446 19 12 to to PART fcis-12446 19 13 replace replace VERB fcis-12446 19 14 manual manual ADJ fcis-12446 19 15 monitoring monitoring NOUN fcis-12446 19 16 is be AUX fcis-12446 19 17 of of ADP fcis-12446 19 18 great great ADJ fcis-12446 19 19 research research NOUN fcis-12446 19 20 and and CCONJ fcis-12446 19 21 practical practical ADJ fcis-12446 19 22 value value NOUN fcis-12446 19 23 . . PUNCT fcis-12446 20 1 smart smart ADJ fcis-12446 20 2 agriculture agriculture NOUN fcis-12446 20 3 is be AUX fcis-12446 20 4 an an DET fcis-12446 20 5 advanced advanced ADJ fcis-12446 20 6 solution solution NOUN fcis-12446 20 7 that that PRON fcis-12446 20 8 uses use VERB fcis-12446 20 9 modern modern ADJ fcis-12446 20 10 information information NOUN fcis-12446 20 11 technology technology NOUN fcis-12446 20 12 to to PART fcis-12446 20 13 control control VERB fcis-12446 20 14 the the DET fcis-12446 20 15 production production NOUN fcis-12446 20 16 process process NOUN fcis-12446 20 17 of of ADP fcis-12446 20 18 agriculture agriculture NOUN fcis-12446 20 19 , , PUNCT fcis-12446 20 20 and and CCONJ fcis-12446 20 21 one one NUM fcis-12446 20 22 of of ADP fcis-12446 20 23 its its PRON fcis-12446 20 24 main main ADJ fcis-12446 20 25 applications application NOUN fcis-12446 20 26 is be AUX fcis-12446 20 27 the the DET fcis-12446 20 28 automated automate VERB fcis-12446 20 29 identification identification NOUN fcis-12446 20 30 of of ADP fcis-12446 20 31 pest pest NOUN fcis-12446 20 32 targets target NOUN fcis-12446 20 33 using use VERB fcis-12446 20 34 artificial artificial ADJ fcis-12446 20 35 intelligence intelligence NOUN fcis-12446 20 36 methods method NOUN fcis-12446 20 37 . . PUNCT fcis-12446 21 1 specifically specifically ADV fcis-12446 21 2 , , PUNCT fcis-12446 21 3 the the DET fcis-12446 21 4 automated automate VERB fcis-12446 21 5 identification identification NOUN fcis-12446 21 6 of of ADP fcis-12446 21 7 pests pest NOUN fcis-12446 21 8 utilizes utilize VERB fcis-12446 21 9 convolutional convolutional ADJ fcis-12446 21 10 neural neural ADJ fcis-12446 21 11 networks network NOUN fcis-12446 21 12 in in ADP fcis-12446 21 13 deep deep ADJ fcis-12446 21 14 learning learning NOUN fcis-12446 21 15 to to PART fcis-12446 21 16 extract extract VERB fcis-12446 21 17 information information NOUN fcis-12446 21 18 from from ADP fcis-12446 21 19 image image NOUN fcis-12446 21 20 data datum NOUN fcis-12446 21 21 , , PUNCT fcis-12446 21 22 then then ADV fcis-12446 21 23 iteratively iteratively ADV fcis-12446 21 24 searches search VERB fcis-12446 21 25 for for ADP fcis-12446 21 26 the the DET fcis-12446 21 27 position position NOUN fcis-12446 21 28 of of ADP fcis-12446 21 29 the the DET fcis-12446 21 30 target target NOUN fcis-12446 21 31 in in ADP fcis-12446 21 32 the the DET fcis-12446 21 33 images image NOUN fcis-12446 21 34 on on ADP fcis-12446 21 35 computer computer NOUN fcis-12446 21 36 devices device NOUN fcis-12446 21 37 and and CCONJ fcis-12446 21 38 determines determine VERB fcis-12446 21 39 the the DET fcis-12446 21 40 target target NOUN fcis-12446 21 41 category category NOUN fcis-12446 21 42 . . PUNCT fcis-12446 22 1 to to PART fcis-12446 22 2 advance advance VERB fcis-12446 22 3 smart smart ADJ fcis-12446 22 4 agriculture agriculture NOUN fcis-12446 22 5 , , PUNCT fcis-12446 22 6 many many ADJ fcis-12446 22 7 scholars scholar NOUN fcis-12446 22 8 have have AUX fcis-12446 22 9 developed develop VERB fcis-12446 22 10 high high ADJ fcis-12446 22 11 - - PUNCT fcis-12446 22 12 performance performance NOUN fcis-12446 22 13 target target NOUN fcis-12446 22 14 detection detection NOUN fcis-12446 22 15 solutions solution NOUN fcis-12446 22 16 for for ADP fcis-12446 22 17 different different ADJ fcis-12446 22 18 pests pest NOUN fcis-12446 22 19 , , PUNCT fcis-12446 22 20 some some PRON fcis-12446 22 21 of of ADP fcis-12446 22 22 which which PRON fcis-12446 22 23 can can AUX fcis-12446 22 24 even even ADV fcis-12446 22 25 achieve achieve VERB fcis-12446 22 26 100 100 NUM fcis-12446 22 27 % % NOUN fcis-12446 22 28 detection detection NOUN fcis-12446 22 29 accuracy accuracy NOUN fcis-12446 22 30 in in ADP fcis-12446 22 31 laboratory laboratory NOUN fcis-12446 22 32 environments environment NOUN fcis-12446 22 33 . . PUNCT fcis-12446 23 1 however however ADV fcis-12446 23 2 , , PUNCT fcis-12446 23 3 applying apply VERB fcis-12446 23 4 target target NOUN fcis-12446 23 5 detection detection NOUN fcis-12446 23 6 solutions solution NOUN fcis-12446 23 7 to to ADP fcis-12446 23 8 practical practical ADJ fcis-12446 23 9 field field NOUN fcis-12446 23 10 pest pest NOUN fcis-12446 23 11 recognition recognition NOUN fcis-12446 23 12 requires require VERB fcis-12446 23 13 considering consider VERB fcis-12446 23 14 more more ADJ fcis-12446 23 15 factors factor NOUN fcis-12446 23 16 such such ADJ fcis-12446 23 17 as as ADP fcis-12446 23 18 lighting lighting NOUN fcis-12446 23 19 and and CCONJ fcis-12446 23 20 occlusion occlusion NOUN fcis-12446 23 21 . . PUNCT fcis-12446 24 1 in in ADP fcis-12446 24 2 addition addition NOUN fcis-12446 24 3 , , PUNCT fcis-12446 24 4 for for ADP fcis-12446 24 5 some some DET fcis-12446 24 6 very very ADV fcis-12446 24 7 small small ADJ fcis-12446 24 8 pests pest NOUN fcis-12446 24 9 in in ADP fcis-12446 24 10 the the DET fcis-12446 24 11 pest pest NOUN fcis-12446 24 12 domain domain NOUN fcis-12446 24 13 , , PUNCT fcis-12446 24 14 such such ADJ fcis-12446 24 15 as as ADP fcis-12446 24 16 aphid aphid NOUN fcis-12446 24 17 targets target NOUN fcis-12446 24 18 commonly commonly ADV fcis-12446 24 19 found find VERB fcis-12446 24 20 on on ADP fcis-12446 24 21 major major ADJ fcis-12446 24 22 crops crop NOUN fcis-12446 24 23 such such ADJ fcis-12446 24 24 as as ADP fcis-12446 24 25 corn corn NOUN fcis-12446 24 26 , , PUNCT fcis-12446 24 27 wheat wheat NOUN fcis-12446 24 28 , , PUNCT fcis-12446 24 29 and and CCONJ fcis-12446 24 30 canola canola NOUN fcis-12446 24 31 , , PUNCT fcis-12446 24 32 there there PRON fcis-12446 24 33 are be VERB fcis-12446 24 34 insufficient insufficient ADJ fcis-12446 24 35 accuracy accuracy NOUN fcis-12446 24 36 and and CCONJ fcis-12446 24 37 recall recall NOUN fcis-12446 24 38 rates rate NOUN fcis-12446 24 39 in in ADP fcis-12446 24 40 detection detection NOUN fcis-12446 24 41 . . PUNCT fcis-12446 25 1 in in ADP fcis-12446 25 2 aphid aphid NOUN fcis-12446 25 3 detection detection NOUN fcis-12446 25 4 , , PUNCT fcis-12446 25 5 there there PRON fcis-12446 25 6 are be VERB fcis-12446 25 7 mainly mainly ADV fcis-12446 25 8 two two NUM fcis-12446 25 9 problems problem NOUN fcis-12446 25 10 . . PUNCT fcis-12446 26 1 on on ADP fcis-12446 26 2 the the DET fcis-12446 26 3 one one NUM fcis-12446 26 4 hand hand NOUN fcis-12446 26 5 , , PUNCT fcis-12446 26 6 aphid aphid NOUN fcis-12446 26 7 targets target NOUN fcis-12446 26 8 are be AUX fcis-12446 26 9 smaller small ADJ fcis-12446 26 10 than than ADP fcis-12446 26 11 most most ADJ fcis-12446 26 12 other other ADJ fcis-12446 26 13 pest pest NOUN fcis-12446 26 14 targets target NOUN fcis-12446 26 15 , , PUNCT fcis-12446 26 16 which which PRON fcis-12446 26 17 makes make VERB fcis-12446 26 18 it it PRON fcis-12446 26 19 more more ADV fcis-12446 26 20 difficult difficult ADJ fcis-12446 26 21 to to PART fcis-12446 26 22 retain retain VERB fcis-12446 26 23 aphid aphid NOUN fcis-12446 26 24 target target NOUN fcis-12446 26 25 information information NOUN fcis-12446 26 26 during during ADP fcis-12446 26 27 feature feature NOUN fcis-12446 26 28 extraction extraction NOUN fcis-12446 26 29 , , PUNCT fcis-12446 26 30 as as SCONJ fcis-12446 26 31 shown show VERB fcis-12446 26 32 in in ADP fcis-12446 26 33 figure figure NOUN fcis-12446 26 34 1 1 NUM fcis-12446 26 35 . . PUNCT fcis-12446 27 1 the the DET fcis-12446 27 2 figure figure NOUN fcis-12446 27 3 shows show VERB fcis-12446 27 4 the the DET fcis-12446 27 5 conventional conventional ADJ fcis-12446 27 6 - - PUNCT fcis-12446 27 7 sized sized ADJ fcis-12446 27 8 pests pest NOUN fcis-12446 27 9 and and CCONJ fcis-12446 27 10 aphid aphid NOUN fcis-12446 27 11 targets target NOUN fcis-12446 27 12 on on ADP fcis-12446 27 13 three three NUM fcis-12446 27 14 common common ADJ fcis-12446 27 15 crops crop NOUN fcis-12446 27 16 : : PUNCT fcis-12446 27 17 corn corn NOUN fcis-12446 27 18 , , PUNCT fcis-12446 27 19 canola canola NOUN fcis-12446 27 20 , , PUNCT fcis-12446 27 21 and and CCONJ fcis-12446 27 22 wheat wheat NOUN fcis-12446 27 23 . . PUNCT fcis-12446 28 1 for for ADP fcis-12446 28 2 example example NOUN fcis-12446 28 3 , , PUNCT fcis-12446 28 4 in in ADP fcis-12446 28 5 the the DET fcis-12446 28 6 case case NOUN fcis-12446 28 7 of of ADP fcis-12446 28 8 corn corn NOUN fcis-12446 28 9 pests pest NOUN fcis-12446 28 10 , , PUNCT fcis-12446 28 11 the the DET fcis-12446 28 12 length length NOUN fcis-12446 28 13 of of ADP fcis-12446 28 14 the the DET fcis-12446 28 15 sugar sugar NOUN fcis-12446 28 16 beet beet PROPN fcis-12446 28 17 night night NOUN fcis-12446 28 18 larva larva NOUN fcis-12446 28 19 is be AUX fcis-12446 28 20 approximately approximately ADV fcis-12446 28 21 30 30 NUM fcis-12446 28 22 mm mm NOUN fcis-12446 28 23 , , PUNCT fcis-12446 28 24 while while SCONJ fcis-12446 28 25 the the DET fcis-12446 28 26 length length NOUN fcis-12446 28 27 of of ADP fcis-12446 28 28 the the DET fcis-12446 28 29 corn corn NOUN fcis-12446 28 30 aphid aphid NOUN fcis-12446 28 31 adult adult NOUN fcis-12446 28 32 is be AUX fcis-12446 28 33 about about ADV fcis-12446 28 34 2.2 2.2 NUM fcis-12446 28 35 mm mm NOUN fcis-12446 28 36 , , PUNCT fcis-12446 28 37 resulting result VERB fcis-12446 28 38 in in ADP fcis-12446 28 39 a a DET fcis-12446 28 40 huge huge ADJ fcis-12446 28 41 difference difference NOUN fcis-12446 28 42 in in ADP fcis-12446 28 43 target target NOUN fcis-12446 28 44 size size NOUN fcis-12446 28 45 , , PUNCT fcis-12446 28 46 which which PRON fcis-12446 28 47 makes make VERB fcis-12446 28 48 it it PRON fcis-12446 28 49 easy easy ADJ fcis-12446 28 50 to to PART fcis-12446 28 51 lose lose VERB fcis-12446 28 52 information information NOUN fcis-12446 28 53 during during ADP fcis-12446 28 54 model model NOUN fcis-12446 28 55 training training NOUN fcis-12446 28 56 . . PUNCT fcis-12446 29 1 on on ADP fcis-12446 29 2 the the DET fcis-12446 29 3 other other ADJ fcis-12446 29 4 hand hand NOUN fcis-12446 29 5 , , PUNCT fcis-12446 29 6 aphid aphid NOUN fcis-12446 29 7 targets target NOUN fcis-12446 29 8 have have VERB fcis-12446 29 9 the the DET fcis-12446 29 10 characteristic characteristic NOUN fcis-12446 29 11 of of ADP fcis-12446 29 12 dense dense ADJ fcis-12446 29 13 distribution distribution NOUN fcis-12446 29 14 , , PUNCT fcis-12446 29 15 which which PRON fcis-12446 29 16 poses pose VERB fcis-12446 29 17 problems problem NOUN fcis-12446 29 18 in in ADP fcis-12446 29 19 pest pest NOUN fcis-12446 29 20 detection detection NOUN fcis-12446 29 21 , , PUNCT fcis-12446 29 22 such such ADJ fcis-12446 29 23 as as ADP fcis-12446 29 24 occlusion occlusion NOUN fcis-12446 29 25 of of ADP fcis-12446 29 26 targets target NOUN fcis-12446 29 27 , , PUNCT fcis-12446 29 28 as as SCONJ fcis-12446 29 29 shown show VERB fcis-12446 29 30 in in ADP fcis-12446 29 31 figure figure NOUN fcis-12446 29 32 2 2 NUM fcis-12446 29 33 . . PUNCT fcis-12446 30 1 the the DET fcis-12446 30 2 figure figure NOUN fcis-12446 30 3 shows show VERB fcis-12446 30 4 the the DET fcis-12446 30 5 distribution distribution NOUN fcis-12446 30 6 of of ADP fcis-12446 30 7 aphid aphid NOUN fcis-12446 30 8 targets target NOUN fcis-12446 30 9 on on ADP fcis-12446 30 10 three three NUM fcis-12446 30 11 crops crop NOUN fcis-12446 30 12 , , PUNCT fcis-12446 30 13 and and CCONJ fcis-12446 30 14 when when SCONJ fcis-12446 30 15 performing perform VERB fcis-12446 30 16 detection detection NOUN fcis-12446 30 17 , , PUNCT fcis-12446 30 18 the the DET fcis-12446 30 19 obscured obscure VERB fcis-12446 30 20 targets target NOUN fcis-12446 30 21 are be AUX fcis-12446 30 22 easily easily ADV fcis-12446 30 23 mistaken mistaken ADJ fcis-12446 30 24 for for ADP fcis-12446 30 25 background background NOUN fcis-12446 30 26 areas area NOUN fcis-12446 30 27 , , PUNCT fcis-12446 30 28 leading lead VERB fcis-12446 30 29 to to ADP fcis-12446 30 30 a a DET fcis-12446 30 31 high high ADJ fcis-12446 30 32 false false ADJ fcis-12446 30 33 negative negative ADJ fcis-12446 30 34 rate rate NOUN fcis-12446 30 35 and and CCONJ fcis-12446 30 36 poor poor ADJ fcis-12446 30 37 model model NOUN fcis-12446 30 38 generalization generalization NOUN fcis-12446 30 39 ability ability NOUN fcis-12446 30 40 . . PUNCT fcis-12446 31 1 2 2 X fcis-12446 31 2 . . X fcis-12446 31 3 related relate VERB fcis-12446 31 4 work work NOUN fcis-12446 31 5 with with ADP fcis-12446 31 6 the the DET fcis-12446 31 7 development development NOUN fcis-12446 31 8 of of ADP fcis-12446 31 9 deep deep ADJ fcis-12446 31 10 learning learning NOUN fcis-12446 31 11 , , PUNCT fcis-12446 31 12 computer computer NOUN fcis-12446 31 13 vision vision NOUN fcis-12446 31 14 capabilities capability NOUN fcis-12446 31 15 have have AUX fcis-12446 31 16 gradually gradually ADV fcis-12446 31 17 enriched enrich VERB fcis-12446 31 18 and and CCONJ fcis-12446 31 19 improved improve VERB fcis-12446 31 20 . . PUNCT fcis-12446 32 1 however however ADV fcis-12446 32 2 , , PUNCT fcis-12446 32 3 in in ADP fcis-12446 32 4 the the DET fcis-12446 32 5 early early ADJ fcis-12446 32 6 stages stage NOUN fcis-12446 32 7 of of ADP fcis-12446 32 8 computer computer NOUN fcis-12446 32 9 vision vision NOUN fcis-12446 32 10 growth growth NOUN fcis-12446 32 11 , , PUNCT fcis-12446 32 12 the the DET fcis-12446 32 13 functionality functionality NOUN fcis-12446 32 14 was be AUX fcis-12446 32 15 somewhat somewhat ADV fcis-12446 32 16 limited limited ADJ fcis-12446 32 17 . . PUNCT fcis-12446 33 1 before before ADP fcis-12446 33 2 the the DET fcis-12446 33 3 introduction introduction NOUN fcis-12446 33 4 of of ADP fcis-12446 33 5 the the DET fcis-12446 33 6 r r NOUN fcis-12446 33 7 - - PUNCT fcis-12446 33 8 cnn cnn NOUN fcis-12446 33 9 [ [ X fcis-12446 33 10 3 3 NUM fcis-12446 33 11 ] ] X fcis-12446 33 12 network network NOUN fcis-12446 33 13 , , PUNCT fcis-12446 33 14 the the DET fcis-12446 33 15 primary primary ADJ fcis-12446 33 16 function function NOUN fcis-12446 33 17 of of ADP fcis-12446 33 18 computer computer NOUN fcis-12446 33 19 vision vision NOUN fcis-12446 33 20 was be AUX fcis-12446 33 21 image image NOUN fcis-12446 33 22 classification classification NOUN fcis-12446 33 23 . . PUNCT fcis-12446 34 1 image image NOUN fcis-12446 34 2 classification classification NOUN fcis-12446 34 3 generally generally ADV fcis-12446 34 4 involves involve VERB fcis-12446 34 5 two two NUM fcis-12446 34 6 processes process NOUN fcis-12446 34 7 : : PUNCT fcis-12446 34 8 first first ADV fcis-12446 34 9 , , PUNCT fcis-12446 34 10 manually manually ADV fcis-12446 34 11 designed design VERB fcis-12446 34 12 feature feature NOUN fcis-12446 34 13 representations representation NOUN fcis-12446 34 14 are be AUX fcis-12446 34 15 extracted extract VERB fcis-12446 34 16 , , PUNCT fcis-12446 34 17 and and CCONJ fcis-12446 34 18 then then ADV fcis-12446 34 19 these these DET fcis-12446 34 20 representations representation NOUN fcis-12446 34 21 , , PUNCT fcis-12446 34 22 which which PRON fcis-12446 34 23 capture capture VERB fcis-12446 34 24 abstract abstract ADJ fcis-12446 34 25 information information NOUN fcis-12446 34 26 about about ADP fcis-12446 34 27 the the DET fcis-12446 34 28 62 62 NUM fcis-12446 34 29 images image NOUN fcis-12446 34 30 , , PUNCT fcis-12446 34 31 are be AUX fcis-12446 34 32 classified classify VERB fcis-12446 34 33 using use VERB fcis-12446 34 34 a a DET fcis-12446 34 35 trained train VERB fcis-12446 34 36 classifier classifier NOUN fcis-12446 34 37 to to PART fcis-12446 34 38 determine determine VERB fcis-12446 34 39 the the DET fcis-12446 34 40 image image NOUN fcis-12446 34 41 category category NOUN fcis-12446 34 42 . . PUNCT fcis-12446 35 1 in in ADP fcis-12446 35 2 the the DET fcis-12446 35 3 early early ADJ fcis-12446 35 4 stages stage NOUN fcis-12446 35 5 of of ADP fcis-12446 35 6 using use VERB fcis-12446 35 7 computer computer NOUN fcis-12446 35 8 vision vision NOUN fcis-12446 35 9 to to PART fcis-12446 35 10 recognize recognize VERB fcis-12446 35 11 pests pest NOUN fcis-12446 35 12 , , PUNCT fcis-12446 35 13 the the DET fcis-12446 35 14 task task NOUN fcis-12446 35 15 was be AUX fcis-12446 35 16 image image NOUN fcis-12446 35 17 classification classification NOUN fcis-12446 35 18 . . PUNCT fcis-12446 36 1 in in ADP fcis-12446 36 2 this this DET fcis-12446 36 3 article article NOUN fcis-12446 36 4 , , PUNCT fcis-12446 36 5 traditional traditional ADJ fcis-12446 36 6 pest pest NOUN fcis-12446 36 7 detection detection NOUN fcis-12446 36 8 is be AUX fcis-12446 36 9 divided divide VERB fcis-12446 36 10 into into ADP fcis-12446 36 11 two two NUM fcis-12446 36 12 categories category NOUN fcis-12446 36 13 based base VERB fcis-12446 36 14 on on ADP fcis-12446 36 15 the the DET fcis-12446 36 16 detection detection NOUN fcis-12446 36 17 scene scene NOUN fcis-12446 36 18 of of ADP fcis-12446 36 19 the the DET fcis-12446 36 20 target target NOUN fcis-12446 36 21 : : PUNCT fcis-12446 36 22 detection detection NOUN fcis-12446 36 23 in in ADP fcis-12446 36 24 experimental experimental ADJ fcis-12446 36 25 environments environment NOUN fcis-12446 36 26 and and CCONJ fcis-12446 36 27 detection detection NOUN fcis-12446 36 28 in in ADP fcis-12446 36 29 natural natural ADJ fcis-12446 36 30 environments environment NOUN fcis-12446 36 31 . . PUNCT fcis-12446 37 1 ( ( PUNCT fcis-12446 37 2 a a X fcis-12446 37 3 ) ) PUNCT fcis-12446 37 4 spodoptera spodoptera NOUN fcis-12446 37 5 exigua exigua NOUN fcis-12446 37 6 ( ( PUNCT fcis-12446 37 7 b b NOUN fcis-12446 37 8 ) ) PUNCT fcis-12446 37 9 rhopalosiphum rhopalosiphum NOUN fcis-12446 37 10 maidis maidis PROPN fcis-12446 37 11 ( ( PUNCT fcis-12446 37 12 c c NOUN fcis-12446 37 13 ) ) PUNCT fcis-12446 37 14 cabbage cabbage NOUN fcis-12446 37 15 butterfly butterfly NOUN fcis-12446 37 16 ( ( PUNCT fcis-12446 37 17 d d NOUN fcis-12446 37 18 ) ) PUNCT fcis-12446 37 19 eristalinae eristalinae NOUN fcis-12446 37 20 ( ( PUNCT fcis-12446 37 21 e e NOUN fcis-12446 37 22 ) ) PUNCT fcis-12446 37 23 mythimna mythimna NOUN fcis-12446 37 24 separta separta PROPN fcis-12446 37 25 ( ( PUNCT fcis-12446 37 26 f f X fcis-12446 37 27 ) ) PUNCT fcis-12446 37 28 macrosiphum macrosiphum NOUN fcis-12446 37 29 avenae avenae NOUN fcis-12446 37 30 figure figure NOUN fcis-12446 37 31 1 1 NUM fcis-12446 37 32 . . PUNCT fcis-12446 37 33 comparison comparison NOUN fcis-12446 37 34 chart chart NOUN fcis-12446 37 35 of of ADP fcis-12446 37 36 pest pest NOUN fcis-12446 37 37 sizes size NOUN fcis-12446 37 38 figure figure VERB fcis-12446 37 39 2 2 NUM fcis-12446 37 40 . . PUNCT fcis-12446 37 41 densely densely ADV fcis-12446 37 42 distributed distribute VERB fcis-12446 37 43 small small ADJ fcis-12446 37 44 target target NOUN fcis-12446 37 45 aphids aphid NOUN fcis-12446 37 46 detection detection NOUN fcis-12446 37 47 in in ADP fcis-12446 37 48 experimental experimental ADJ fcis-12446 37 49 environments environment NOUN fcis-12446 37 50 refers refer VERB fcis-12446 37 51 to to ADP fcis-12446 37 52 capturing capture VERB fcis-12446 37 53 field field NOUN fcis-12446 37 54 pests pest NOUN fcis-12446 37 55 by by ADP fcis-12446 37 56 designing design VERB fcis-12446 37 57 attractors attractor NOUN fcis-12446 37 58 through through ADP fcis-12446 37 59 adjustments adjustment NOUN fcis-12446 37 60 such such ADJ fcis-12446 37 61 as as ADP fcis-12446 37 62 lighting lighting NOUN fcis-12446 37 63 and and CCONJ fcis-12446 37 64 placing place VERB fcis-12446 37 65 food food NOUN fcis-12446 37 66 . . PUNCT fcis-12446 38 1 then then ADV fcis-12446 38 2 , , PUNCT fcis-12446 38 3 image image NOUN fcis-12446 38 4 information information NOUN fcis-12446 38 5 is be AUX fcis-12446 38 6 collected collect VERB fcis-12446 38 7 to to PART fcis-12446 38 8 create create VERB fcis-12446 38 9 an an DET fcis-12446 38 10 experimental experimental ADJ fcis-12446 38 11 dataset dataset NOUN fcis-12446 38 12 . . PUNCT fcis-12446 39 1 finally finally ADV fcis-12446 39 2 , , PUNCT fcis-12446 39 3 a a DET fcis-12446 39 4 classifier classifier NOUN fcis-12446 39 5 is be AUX fcis-12446 39 6 trained train VERB fcis-12446 39 7 to to PART fcis-12446 39 8 achieve achieve VERB fcis-12446 39 9 artificial artificial ADJ fcis-12446 39 10 intelligence intelligence NOUN fcis-12446 39 11 classification classification NOUN fcis-12446 39 12 of of ADP fcis-12446 39 13 the the DET fcis-12446 39 14 images image NOUN fcis-12446 39 15 in in ADP fcis-12446 39 16 the the DET fcis-12446 39 17 dataset dataset NOUN fcis-12446 39 18 . . PUNCT fcis-12446 40 1 the the DET fcis-12446 40 2 image image NOUN fcis-12446 40 3 data datum NOUN fcis-12446 40 4 collected collect VERB fcis-12446 40 5 in in ADP fcis-12446 40 6 this this DET fcis-12446 40 7 way way NOUN fcis-12446 40 8 is be AUX fcis-12446 40 9 non non ADJ fcis-12446 40 10 - - ADJ fcis-12446 40 11 natural natural ADJ fcis-12446 40 12 and and CCONJ fcis-12446 40 13 difficult difficult ADJ fcis-12446 40 14 to to PART fcis-12446 40 15 be be AUX fcis-12446 40 16 used use VERB fcis-12446 40 17 for for ADP fcis-12446 40 18 practical practical ADJ fcis-12446 40 19 detection detection NOUN fcis-12446 40 20 in in ADP fcis-12446 40 21 complex complex ADJ fcis-12446 40 22 agricultural agricultural ADJ fcis-12446 40 23 environments environment NOUN fcis-12446 40 24 . . PUNCT fcis-12446 41 1 shahrul shahrul NOUN fcis-12446 41 2 et et PROPN fcis-12446 41 3 al al PROPN fcis-12446 41 4 . . PUNCT fcis-12446 42 1 [ [ X fcis-12446 42 2 4 4 X fcis-12446 42 3 ] ] PUNCT fcis-12446 42 4 used use VERB fcis-12446 42 5 quality quality NOUN fcis-12446 42 6 thresholds threshold NOUN fcis-12446 42 7 and and CCONJ fcis-12446 42 8 moment moment NOUN fcis-12446 42 9 invariants invariant VERB fcis-12446 42 10 to to PART fcis-12446 42 11 extract extract VERB fcis-12446 42 12 features feature NOUN fcis-12446 42 13 of of ADP fcis-12446 42 14 20 20 NUM fcis-12446 42 15 different different ADJ fcis-12446 42 16 insects insect NOUN fcis-12446 42 17 for for ADP fcis-12446 42 18 classifier classifier NOUN fcis-12446 42 19 training training NOUN fcis-12446 42 20 and and CCONJ fcis-12446 42 21 introduced introduce VERB fcis-12446 42 22 intra intra ADJ fcis-12446 42 23 - - ADJ fcis-12446 42 24 class class ADJ fcis-12446 42 25 analysis analysis NOUN fcis-12446 42 26 methods method NOUN fcis-12446 42 27 to to PART fcis-12446 42 28 evaluate evaluate VERB fcis-12446 42 29 the the DET fcis-12446 42 30 position position NOUN fcis-12446 42 31 sensitivity sensitivity NOUN fcis-12446 42 32 and and CCONJ fcis-12446 42 33 rotation rotation NOUN fcis-12446 42 34 invariance invariance NOUN fcis-12446 42 35 of of ADP fcis-12446 42 36 images image NOUN fcis-12446 42 37 . . PUNCT fcis-12446 43 1 wang wang PROPN fcis-12446 43 2 et et PROPN fcis-12446 43 3 al al PROPN fcis-12446 43 4 . . PUNCT fcis-12446 44 1 [ [ X fcis-12446 44 2 5 5 NUM fcis-12446 44 3 ] ] PUNCT fcis-12446 44 4 proposed propose VERB fcis-12446 44 5 an an DET fcis-12446 44 6 insect insect NOUN fcis-12446 44 7 recognition recognition NOUN fcis-12446 44 8 system system NOUN fcis-12446 44 9 based base VERB fcis-12446 44 10 on on ADP fcis-12446 44 11 the the DET fcis-12446 44 12 family family NOUN fcis-12446 44 13 level level NOUN fcis-12446 44 14 , , PUNCT fcis-12446 44 15 classifying classify VERB fcis-12446 44 16 and and CCONJ fcis-12446 44 17 identifying identify VERB fcis-12446 44 18 225 225 NUM fcis-12446 44 19 images image NOUN fcis-12446 44 20 of of ADP fcis-12446 44 21 9 9 NUM fcis-12446 44 22 common common ADJ fcis-12446 44 23 insect insect NOUN fcis-12446 44 24 orders order NOUN fcis-12446 44 25 . . PUNCT fcis-12446 45 1 with with ADP fcis-12446 45 2 the the DET fcis-12446 45 3 support support NOUN fcis-12446 45 4 of of ADP fcis-12446 45 5 artificial artificial ADJ fcis-12446 45 6 neural neural ADJ fcis-12446 45 7 networks network NOUN fcis-12446 45 8 , , PUNCT fcis-12446 45 9 they they PRON fcis-12446 45 10 achieved achieve VERB fcis-12446 45 11 a a DET fcis-12446 45 12 93 93 NUM fcis-12446 45 13 % % NOUN fcis-12446 45 14 accuracy accuracy NOUN fcis-12446 45 15 rate rate NOUN fcis-12446 45 16 . . PUNCT fcis-12446 46 1 to to PART fcis-12446 46 2 reduce reduce VERB fcis-12446 46 3 the the DET fcis-12446 46 4 reliance reliance NOUN fcis-12446 46 5 on on ADP fcis-12446 46 6 human human ADJ fcis-12446 46 7 efforts effort NOUN fcis-12446 46 8 for for ADP fcis-12446 46 9 pest pest NOUN fcis-12446 46 10 recognition recognition NOUN fcis-12446 46 11 , , PUNCT fcis-12446 46 12 qing qe VERB fcis-12446 46 13 et et PROPN fcis-12446 46 14 al al PROPN fcis-12446 46 15 . . PUNCT fcis-12446 47 1 [ [ X fcis-12446 47 2 6 6 NUM fcis-12446 47 3 ] ] PUNCT fcis-12446 47 4 designed design VERB fcis-12446 47 5 an an DET fcis-12446 47 6 automated automate VERB fcis-12446 47 7 imaging imaging NOUN fcis-12446 47 8 system system NOUN fcis-12446 47 9 for for ADP fcis-12446 47 10 identifying identify VERB fcis-12446 47 11 light light ADV fcis-12446 47 12 - - PUNCT fcis-12446 47 13 induced induce VERB fcis-12446 47 14 rice rice NOUN fcis-12446 47 15 field field NOUN fcis-12446 47 16 pests pest NOUN fcis-12446 47 17 . . PUNCT fcis-12446 48 1 they they PRON fcis-12446 48 2 extracted extract VERB fcis-12446 48 3 156 156 NUM fcis-12446 48 4 features feature NOUN fcis-12446 48 5 including include VERB fcis-12446 48 6 color color NOUN fcis-12446 48 7 and and CCONJ fcis-12446 48 8 shape shape NOUN fcis-12446 48 9 and and CCONJ fcis-12446 48 10 trained train VERB fcis-12446 48 11 a a DET fcis-12446 48 12 support support NOUN fcis-12446 48 13 vector vector NOUN fcis-12446 48 14 machine machine NOUN fcis-12446 48 15 for for ADP fcis-12446 48 16 classification classification NOUN fcis-12446 48 17 , , PUNCT fcis-12446 48 18 achieving achieve VERB fcis-12446 48 19 a a DET fcis-12446 48 20 97.5 97.5 NUM fcis-12446 48 21 % % NOUN fcis-12446 48 22 accuracy accuracy NOUN fcis-12446 48 23 rate rate NOUN fcis-12446 48 24 . . PUNCT fcis-12446 49 1 for for ADP fcis-12446 49 2 6 6 NUM fcis-12446 49 3 common common ADJ fcis-12446 49 4 field field NOUN fcis-12446 49 5 pests pest NOUN fcis-12446 49 6 , , PUNCT fcis-12446 49 7 han han PROPN fcis-12446 49 8 rui rui PROPN fcis-12446 49 9 - - PUNCT fcis-12446 49 10 zhen zhen PROPN fcis-12446 49 11 et et PROPN fcis-12446 49 12 al al PROPN fcis-12446 49 13 . . PUNCT fcis-12446 50 1 [ [ X fcis-12446 50 2 7 7 X fcis-12446 50 3 ] ] PUNCT fcis-12446 50 4 extracted extract VERB fcis-12446 50 5 25 25 NUM fcis-12446 50 6 features feature NOUN fcis-12446 50 7 based base VERB fcis-12446 50 8 on on ADP fcis-12446 50 9 morphology morphology NOUN fcis-12446 50 10 and and CCONJ fcis-12446 50 11 color color NOUN fcis-12446 50 12 and and CCONJ fcis-12446 50 13 trained train VERB fcis-12446 50 14 a a DET fcis-12446 50 15 support support NOUN fcis-12446 50 16 vector vector NOUN fcis-12446 50 17 machine machine NOUN fcis-12446 50 18 classifier classifier NOUN fcis-12446 50 19 , , PUNCT fcis-12446 50 20 ultimately ultimately ADV fcis-12446 50 21 achieving achieve VERB fcis-12446 50 22 an an DET fcis-12446 50 23 87.4 87.4 NUM fcis-12446 50 24 % % NOUN fcis-12446 50 25 accuracy accuracy NOUN fcis-12446 50 26 rate rate NOUN fcis-12446 50 27 , , PUNCT fcis-12446 50 28 providing provide VERB fcis-12446 50 29 new new ADJ fcis-12446 50 30 ideas idea NOUN fcis-12446 50 31 for for ADP fcis-12446 50 32 field field NOUN fcis-12446 50 33 pest pest NOUN fcis-12446 50 34 monitoring monitoring NOUN fcis-12446 50 35 and and CCONJ fcis-12446 50 36 control control NOUN fcis-12446 50 37 at at ADP fcis-12446 50 38 the the DET fcis-12446 50 39 time time NOUN fcis-12446 50 40 . . PUNCT fcis-12446 51 1 li li PROPN fcis-12446 51 2 wen wen PROPN fcis-12446 51 3 - - PROPN fcis-12446 51 4 yong yong PROPN fcis-12446 51 5 et et PROPN fcis-12446 51 6 al al PROPN fcis-12446 51 7 . . PUNCT fcis-12446 52 1 [ [ X fcis-12446 52 2 8 8 NUM fcis-12446 52 3 ] ] PUNCT fcis-12446 52 4 collected collect VERB fcis-12446 52 5 320 320 NUM fcis-12446 52 6 images image NOUN fcis-12446 52 7 of of ADP fcis-12446 52 8 four four NUM fcis-12446 52 9 pests pest NOUN fcis-12446 52 10 , , PUNCT fcis-12446 52 11 including include VERB fcis-12446 52 12 aphids aphid NOUN fcis-12446 52 13 , , PUNCT fcis-12446 52 14 and and CCONJ fcis-12446 52 15 extracted extract VERB fcis-12446 52 16 rgb rgb PROPN fcis-12446 52 17 and and CCONJ fcis-12446 52 18 hsv hsv PROPN fcis-12446 52 19 color color NOUN fcis-12446 52 20 features feature NOUN fcis-12446 52 21 from from ADP fcis-12446 52 22 the the DET fcis-12446 52 23 images image NOUN fcis-12446 52 24 . . PUNCT fcis-12446 53 1 they they PRON fcis-12446 53 2 achieved achieve VERB fcis-12446 53 3 a a DET fcis-12446 53 4 100 100 NUM fcis-12446 53 5 % % NOUN fcis-12446 53 6 recognition recognition NOUN fcis-12446 53 7 accuracy accuracy NOUN fcis-12446 53 8 with with ADP fcis-12446 53 9 a a DET fcis-12446 53 10 multiclass multiclass ADJ fcis-12446 53 11 support support NOUN fcis-12446 53 12 vector vector NOUN fcis-12446 53 13 machine machine NOUN fcis-12446 53 14 . . PUNCT fcis-12446 54 1 although although SCONJ fcis-12446 54 2 these these DET fcis-12446 54 3 methods method NOUN fcis-12446 54 4 achieved achieve VERB fcis-12446 54 5 high high ADJ fcis-12446 54 6 detection detection NOUN fcis-12446 54 7 accuracy accuracy NOUN fcis-12446 54 8 , , PUNCT fcis-12446 54 9 the the DET fcis-12446 54 10 images image NOUN fcis-12446 54 11 used use VERB fcis-12446 54 12 for for ADP fcis-12446 54 13 training training NOUN fcis-12446 54 14 were be AUX fcis-12446 54 15 obtained obtain VERB fcis-12446 54 16 in in ADP fcis-12446 54 17 experimental experimental ADJ fcis-12446 54 18 setups setup NOUN fcis-12446 54 19 such such ADJ fcis-12446 54 20 as as ADP fcis-12446 54 21 attractors attractor NOUN fcis-12446 54 22 , , PUNCT fcis-12446 54 23 which which PRON fcis-12446 54 24 lacked lack VERB fcis-12446 54 25 the the DET fcis-12446 54 26 impact impact NOUN fcis-12446 54 27 of of ADP fcis-12446 54 28 factors factor NOUN fcis-12446 54 29 such such ADJ fcis-12446 54 30 as as ADP fcis-12446 54 31 lighting lighting NOUN fcis-12446 54 32 and and CCONJ fcis-12446 54 33 occlusion occlusion NOUN fcis-12446 54 34 compared compare VERB fcis-12446 54 35 to to ADP fcis-12446 54 36 actual actual ADJ fcis-12446 54 37 field field NOUN fcis-12446 54 38 environments environment NOUN fcis-12446 54 39 . . PUNCT fcis-12446 55 1 additionally additionally ADV fcis-12446 55 2 , , PUNCT fcis-12446 55 3 the the DET fcis-12446 55 4 detection detection NOUN fcis-12446 55 5 methods method NOUN fcis-12446 55 6 were be AUX fcis-12446 55 7 not not PART fcis-12446 55 8 flexible flexible ADJ fcis-12446 55 9 enough enough ADV fcis-12446 55 10 and and CCONJ fcis-12446 55 11 could could AUX fcis-12446 55 12 not not PART fcis-12446 55 13 be be AUX fcis-12446 55 14 applied apply VERB fcis-12446 55 15 in in ADP fcis-12446 55 16 practical practical ADJ fcis-12446 55 17 applications application NOUN fcis-12446 55 18 . . PUNCT fcis-12446 56 1 detection detection NOUN fcis-12446 56 2 in in ADP fcis-12446 56 3 natural natural ADJ fcis-12446 56 4 environments environment NOUN fcis-12446 56 5 refers refer VERB fcis-12446 56 6 to to ADP fcis-12446 56 7 capturing capture VERB fcis-12446 56 8 pest pest NOUN fcis-12446 56 9 images image NOUN fcis-12446 56 10 directly directly ADV fcis-12446 56 11 in in ADP fcis-12446 56 12 the the DET fcis-12446 56 13 field field NOUN fcis-12446 56 14 and and CCONJ fcis-12446 56 15 training training NOUN fcis-12446 56 16 detection detection NOUN fcis-12446 56 17 solutions solution NOUN fcis-12446 56 18 on on ADP fcis-12446 56 19 data datum NOUN fcis-12446 56 20 constructed construct VERB fcis-12446 56 21 through through ADP fcis-12446 56 22 manual manual ADJ fcis-12446 56 23 selection selection NOUN fcis-12446 56 24 and and CCONJ fcis-12446 56 25 annotation annotation NOUN fcis-12446 56 26 . . PUNCT fcis-12446 57 1 weeks week NOUN fcis-12446 57 2 et et PROPN fcis-12446 57 3 al al PROPN fcis-12446 57 4 . . PUNCT fcis-12446 58 1 [ [ X fcis-12446 58 2 9 9 NUM fcis-12446 58 3 ] ] PUNCT fcis-12446 58 4 used use VERB fcis-12446 58 5 principal principal ADJ fcis-12446 58 6 component component NOUN fcis-12446 58 7 analysis analysis NOUN fcis-12446 58 8 to to PART fcis-12446 58 9 construct construct VERB fcis-12446 58 10 a a DET fcis-12446 58 11 classifier classifier NOUN fcis-12446 58 12 for for ADP fcis-12446 58 13 identifying identify VERB fcis-12446 58 14 unknown unknown ADJ fcis-12446 58 15 insect insect NOUN fcis-12446 58 16 specimen speciman NOUN fcis-12446 58 17 images image NOUN fcis-12446 58 18 , , PUNCT fcis-12446 58 19 further far ADV fcis-12446 58 20 demonstrating demonstrate VERB fcis-12446 58 21 the the DET fcis-12446 58 22 tremendous tremendous ADJ fcis-12446 58 23 potential potential NOUN fcis-12446 58 24 of of ADP fcis-12446 58 25 convolutional convolutional ADJ fcis-12446 58 26 neural neural ADJ fcis-12446 58 27 networks network NOUN fcis-12446 58 28 in in ADP fcis-12446 58 29 image image NOUN fcis-12446 58 30 classification classification NOUN fcis-12446 58 31 . . PUNCT fcis-12446 59 1 cho cho PROPN fcis-12446 59 2 j j PROPN fcis-12446 59 3 et et PROPN fcis-12446 59 4 al al PROPN fcis-12446 59 5 . . PUNCT fcis-12446 60 1 [ [ X fcis-12446 60 2 10 10 NUM fcis-12446 60 3 ] ] PUNCT fcis-12446 60 4 designed design VERB fcis-12446 60 5 an an DET fcis-12446 60 6 automatic automatic ADJ fcis-12446 60 7 identification identification NOUN fcis-12446 60 8 and and CCONJ fcis-12446 60 9 detection detection NOUN fcis-12446 60 10 model model NOUN fcis-12446 60 11 based base VERB fcis-12446 60 12 on on ADP fcis-12446 60 13 the the DET fcis-12446 60 14 size size NOUN fcis-12446 60 15 and and CCONJ fcis-12446 60 16 color color NOUN fcis-12446 60 17 features feature NOUN fcis-12446 60 18 of of ADP fcis-12446 60 19 greenhouse greenhouse NOUN fcis-12446 60 20 whiteflies whitefly NOUN fcis-12446 60 21 , , PUNCT fcis-12446 60 22 thrips thrip NOUN fcis-12446 60 23 , , PUNCT fcis-12446 60 24 and and CCONJ fcis-12446 60 25 other other ADJ fcis-12446 60 26 targets target NOUN fcis-12446 60 27 . . PUNCT fcis-12446 61 1 they they PRON fcis-12446 61 2 achieved achieve VERB fcis-12446 61 3 an an DET fcis-12446 61 4 accuracy accuracy NOUN fcis-12446 61 5 rate rate NOUN fcis-12446 61 6 of of ADP fcis-12446 61 7 89.7 89.7 NUM fcis-12446 61 8 % % NOUN fcis-12446 61 9 for for ADP fcis-12446 61 10 six six NUM fcis-12446 61 11 pests pest NOUN fcis-12446 61 12 . . PUNCT fcis-12446 62 1 solis solis PROPN fcis-12446 62 2 et et PROPN fcis-12446 62 3 al al PROPN fcis-12446 62 4 . . PUNCT fcis-12446 63 1 [ [ X fcis-12446 63 2 11 11 NUM fcis-12446 63 3 ] ] PUNCT fcis-12446 63 4 combined combine VERB fcis-12446 63 5 loss loss NOUN fcis-12446 63 6 algorithms algorithm NOUN fcis-12446 63 7 with with ADP fcis-12446 63 8 scale scale NOUN fcis-12446 63 9 - - PUNCT fcis-12446 63 10 invariant invariant ADJ fcis-12446 63 11 features feature NOUN fcis-12446 63 12 to to PART fcis-12446 63 13 propose propose VERB fcis-12446 63 14 a a DET fcis-12446 63 15 new new ADJ fcis-12446 63 16 pest pest NOUN fcis-12446 63 17 detection detection NOUN fcis-12446 63 18 algorithm algorithm NOUN fcis-12446 63 19 , , PUNCT fcis-12446 63 20 which which PRON fcis-12446 63 21 achieved achieve VERB fcis-12446 63 22 an an DET fcis-12446 63 23 average average ADJ fcis-12446 63 24 accuracy accuracy NOUN fcis-12446 63 25 of of ADP fcis-12446 63 26 96.4 96.4 NUM fcis-12446 63 27 % % NOUN fcis-12446 63 28 for for ADP fcis-12446 63 29 six six NUM fcis-12446 63 30 common common ADJ fcis-12446 63 31 pests pest NOUN fcis-12446 63 32 in in ADP fcis-12446 63 33 greenhouses greenhouse NOUN fcis-12446 63 34 . . PUNCT fcis-12446 64 1 they they PRON fcis-12446 64 2 compared compare VERB fcis-12446 64 3 the the DET fcis-12446 64 4 results result NOUN fcis-12446 64 5 with with ADP fcis-12446 64 6 manual manual ADJ fcis-12446 64 7 statistics statistic NOUN fcis-12446 64 8 , , PUNCT fcis-12446 64 9 demonstrating demonstrate VERB fcis-12446 64 10 the the DET fcis-12446 64 11 effectiveness effectiveness NOUN fcis-12446 64 12 of of ADP fcis-12446 64 13 the the DET fcis-12446 64 14 algorithm algorithm NOUN fcis-12446 64 15 . . PUNCT fcis-12446 65 1 although although SCONJ fcis-12446 65 2 these these DET fcis-12446 65 3 methods method NOUN fcis-12446 65 4 are be AUX fcis-12446 65 5 more more ADV fcis-12446 65 6 adaptable adaptable ADJ fcis-12446 65 7 to to ADP fcis-12446 65 8 complex complex ADJ fcis-12446 65 9 environments environment NOUN fcis-12446 65 10 , , PUNCT fcis-12446 65 11 they they PRON fcis-12446 65 12 still still ADV fcis-12446 65 13 suffer suffer VERB fcis-12446 65 14 from from ADP fcis-12446 65 15 drawbacks drawback NOUN fcis-12446 65 16 such such ADJ fcis-12446 65 17 as as ADP fcis-12446 65 18 slow slow ADJ fcis-12446 65 19 detection detection NOUN fcis-12446 65 20 speed speed NOUN fcis-12446 65 21 and and CCONJ fcis-12446 65 22 limited limited ADJ fcis-12446 65 23 detection detection NOUN fcis-12446 65 24 functionality functionality NOUN fcis-12446 65 25 . . PUNCT fcis-12446 66 1 3 3 X fcis-12446 66 2 . . X fcis-12446 66 3 network network NOUN fcis-12446 66 4 proposal proposal NOUN fcis-12446 66 5 to to PART fcis-12446 66 6 address address VERB fcis-12446 66 7 the the DET fcis-12446 66 8 challenges challenge NOUN fcis-12446 66 9 of of ADP fcis-12446 66 10 dense dense ADJ fcis-12446 66 11 distribution distribution NOUN fcis-12446 66 12 and and CCONJ fcis-12446 66 13 small small ADJ fcis-12446 66 14 target target NOUN fcis-12446 66 15 size size NOUN fcis-12446 66 16 in in ADP fcis-12446 66 17 aphid aphid NOUN fcis-12446 66 18 detection detection NOUN fcis-12446 66 19 , , PUNCT fcis-12446 66 20 we we PRON fcis-12446 66 21 made make VERB fcis-12446 66 22 improvements improvement NOUN fcis-12446 66 23 on on ADP fcis-12446 66 24 yolov5 yolov5 NOUN fcis-12446 66 25 , , PUNCT fcis-12446 66 26 as as SCONJ fcis-12446 66 27 shown show VERB fcis-12446 66 28 in in ADP fcis-12446 66 29 figure figure NOUN fcis-12446 66 30 3 3 NUM fcis-12446 66 31 . . NOUN fcis-12446 66 32 63 63 NUM fcis-12446 66 33 feature feature NOUN fcis-12446 66 34     SPACE fcis-12446 66 35 map map NOUN fcis-12446 66 36 h13 h13 NOUN fcis-12446 66 37 h21h11 h21h11 NOUN fcis-12446 66 38 h12 h12 PROPN fcis-12446 66 39 h22 h22 PROPN fcis-12446 66 40 h23 h23 PROPN fcis-12446 66 41 figure figure NOUN fcis-12446 66 42 3 3 NUM fcis-12446 66 43 . . PUNCT fcis-12446 66 44 framework framework NOUN fcis-12446 66 45 diagram diagram NOUN fcis-12446 66 46 of of ADP fcis-12446 66 47 the the DET fcis-12446 66 48 aphid aphid NOUN fcis-12446 66 49 detection detection NOUN fcis-12446 66 50 network network NOUN fcis-12446 66 51 first first ADV fcis-12446 66 52 , , PUNCT fcis-12446 66 53 we we PRON fcis-12446 66 54 extract extract VERB fcis-12446 66 55 features feature NOUN fcis-12446 66 56 from from ADP fcis-12446 66 57 pest pest NOUN fcis-12446 66 58 images image NOUN fcis-12446 66 59 using use VERB fcis-12446 66 60 a a DET fcis-12446 66 61 convolutional convolutional ADJ fcis-12446 66 62 neural neural ADJ fcis-12446 66 63 network network NOUN fcis-12446 66 64 with with ADP fcis-12446 66 65 an an DET fcis-12446 66 66 attention attention NOUN fcis-12446 66 67 mechanism mechanism NOUN fcis-12446 66 68 . . PUNCT fcis-12446 67 1 then then ADV fcis-12446 67 2 , , PUNCT fcis-12446 67 3 we we PRON fcis-12446 67 4 perform perform VERB fcis-12446 67 5 multi multi ADJ fcis-12446 67 6 - - ADJ fcis-12446 67 7 scale scale ADJ fcis-12446 67 8 feature feature NOUN fcis-12446 67 9 fusion fusion NOUN fcis-12446 67 10 on on ADP fcis-12446 67 11 the the DET fcis-12446 67 12 obtained obtain VERB fcis-12446 67 13 feature feature NOUN fcis-12446 67 14 maps map NOUN fcis-12446 67 15 . . PUNCT fcis-12446 68 1 next next ADV fcis-12446 68 2 , , PUNCT fcis-12446 68 3 we we PRON fcis-12446 68 4 feed feed VERB fcis-12446 68 5 the the DET fcis-12446 68 6 three three NUM fcis-12446 68 7 different different ADJ fcis-12446 68 8 scale scale NOUN fcis-12446 68 9 feature feature NOUN fcis-12446 68 10 maps map NOUN fcis-12446 68 11 into into ADP fcis-12446 68 12 the the DET fcis-12446 68 13 detection detection NOUN fcis-12446 68 14 head head NOUN fcis-12446 68 15 for for ADP fcis-12446 68 16 dense dense ADJ fcis-12446 68 17 area area NOUN fcis-12446 68 18 classification classification NOUN fcis-12446 68 19 and and CCONJ fcis-12446 68 20 regression regression NOUN fcis-12446 68 21 . . PUNCT fcis-12446 69 1 finally finally ADV fcis-12446 69 2 , , PUNCT fcis-12446 69 3 we we PRON fcis-12446 69 4 send send VERB fcis-12446 69 5 the the DET fcis-12446 69 6 obtained obtain VERB fcis-12446 69 7 dense dense ADJ fcis-12446 69 8 feature feature NOUN fcis-12446 69 9 regions region NOUN fcis-12446 69 10 to to ADP fcis-12446 69 11 the the DET fcis-12446 69 12 detection detection NOUN fcis-12446 69 13 head head NOUN fcis-12446 69 14 for for ADP fcis-12446 69 15 small small ADJ fcis-12446 69 16 target target NOUN fcis-12446 69 17 pest pest NOUN fcis-12446 69 18 detection detection NOUN fcis-12446 69 19 . . PUNCT fcis-12446 70 1 given give VERB fcis-12446 70 2 that that SCONJ fcis-12446 70 3 most most ADJ fcis-12446 70 4 of of ADP fcis-12446 70 5 the the DET fcis-12446 70 6 target target NOUN fcis-12446 70 7 detection detection NOUN fcis-12446 70 8 frameworks framework NOUN fcis-12446 70 9 proposed propose VERB fcis-12446 70 10 in in ADP fcis-12446 70 11 recent recent ADJ fcis-12446 70 12 years year NOUN fcis-12446 70 13 consist consist VERB fcis-12446 70 14 of of ADP fcis-12446 70 15 a a DET fcis-12446 70 16 backbone backbone NOUN fcis-12446 70 17 network network NOUN fcis-12446 70 18 , , PUNCT fcis-12446 70 19 feature feature NOUN fcis-12446 70 20 fusion fusion NOUN fcis-12446 70 21 modules module NOUN fcis-12446 70 22 , , PUNCT fcis-12446 70 23 and and CCONJ fcis-12446 70 24 detection detection NOUN fcis-12446 70 25 heads head NOUN fcis-12446 70 26 , , PUNCT fcis-12446 70 27 we we PRON fcis-12446 70 28 use use VERB fcis-12446 70 29 this this PRON fcis-12446 70 30 as as ADP fcis-12446 70 31 a a DET fcis-12446 70 32 basis basis NOUN fcis-12446 70 33 to to PART fcis-12446 70 34 provide provide VERB fcis-12446 70 35 a a DET fcis-12446 70 36 detailed detailed ADJ fcis-12446 70 37 description description NOUN fcis-12446 70 38 of of ADP fcis-12446 70 39 the the DET fcis-12446 70 40 improved improved ADJ fcis-12446 70 41 yolov5 yolov5 NOUN fcis-12446 70 42 network network NOUN fcis-12446 70 43 . . PUNCT fcis-12446 71 1 the the DET fcis-12446 71 2 yolov5 yolov5 NOUN fcis-12446 71 3 version version NOUN fcis-12446 71 4 6.1 6.1 NUM fcis-12446 71 5 backbone backbone NOUN fcis-12446 71 6 uses use VERB fcis-12446 71 7 ten ten NUM fcis-12446 71 8 convolutional convolutional ADJ fcis-12446 71 9 modules module NOUN fcis-12446 71 10 for for ADP fcis-12446 71 11 feature feature NOUN fcis-12446 71 12 extraction extraction NOUN fcis-12446 71 13 , , PUNCT fcis-12446 71 14 resulting result VERB fcis-12446 71 15 in in ADP fcis-12446 71 16 three three NUM fcis-12446 71 17 differentscale differentscale ADJ fcis-12446 71 18 feature feature NOUN fcis-12446 71 19 maps map NOUN fcis-12446 71 20 . . PUNCT fcis-12446 72 1 the the DET fcis-12446 72 2 detailed detailed ADJ fcis-12446 72 3 architecture architecture NOUN fcis-12446 72 4 of of ADP fcis-12446 72 5 the the DET fcis-12446 72 6 network network NOUN fcis-12446 72 7 is be AUX fcis-12446 72 8 shown show VERB fcis-12446 72 9 in in ADP fcis-12446 72 10 figure figure NOUN fcis-12446 72 11 4 4 NUM fcis-12446 72 12 : : PUNCT fcis-12446 73 1 conv conv PROPN fcis-12446 73 2 conv conv PROPN fcis-12446 73 3 c3 c3 PROPN fcis-12446 73 4 conv conv PROPN fcis-12446 73 5 c3 c3 PROPN fcis-12446 73 6 conv conv PROPN fcis-12446 73 7 c3 c3 PROPN fcis-12446 73 8 conv conv PROPN fcis-12446 73 9 sppfc3 sppfc3 PROPN fcis-12446 73 10 ta ta X fcis-12446 73 11 ta ta X fcis-12446 73 12 ta ta AUX fcis-12446 73 13 ta ta PART fcis-12446 73 14 figure figure VERB fcis-12446 73 15 4 4 NUM fcis-12446 73 16 . . PUNCT fcis-12446 73 17 backbone backbone NOUN fcis-12446 73 18 network network NOUN fcis-12446 73 19 of of ADP fcis-12446 73 20 yolov5 yolov5 PROPN fcis-12446 73 21 . . PUNCT fcis-12446 74 1 specifically specifically ADV fcis-12446 74 2 , , PUNCT fcis-12446 74 3 the the DET fcis-12446 74 4 backbone backbone NOUN fcis-12446 74 5 of of ADP fcis-12446 74 6 yolov5 yolov5 NOUN fcis-12446 74 7 , , PUNCT fcis-12446 74 8 as as SCONJ fcis-12446 74 9 shown show VERB fcis-12446 74 10 in in ADP fcis-12446 74 11 the the DET fcis-12446 74 12 gray gray ADJ fcis-12446 74 13 part part NOUN fcis-12446 74 14 in in ADP fcis-12446 74 15 figure figure NOUN fcis-12446 74 16 2 2 NUM fcis-12446 74 17 , , PUNCT fcis-12446 74 18 consists consist VERB fcis-12446 74 19 of of ADP fcis-12446 74 20 5 5 NUM fcis-12446 74 21 convolutional convolutional ADJ fcis-12446 74 22 layers layer NOUN fcis-12446 74 23 , , PUNCT fcis-12446 74 24 4 4 NUM fcis-12446 74 25 c3 c3 NOUN fcis-12446 74 26 modules module NOUN fcis-12446 74 27 , , PUNCT fcis-12446 74 28 and and CCONJ fcis-12446 74 29 an an DET fcis-12446 74 30 sppf sppf ADJ fcis-12446 74 31 ( ( PUNCT fcis-12446 74 32 spatial spatial ADJ fcis-12446 74 33 pyramid pyramid NOUN fcis-12446 74 34 pooling pool VERB fcis-12446 74 35 fusion fusion NOUN fcis-12446 74 36 ) ) PUNCT fcis-12446 74 37 layer layer NOUN fcis-12446 74 38 . . PUNCT fcis-12446 75 1 the the DET fcis-12446 75 2 convolutional convolutional ADJ fcis-12446 75 3 layers layer NOUN fcis-12446 75 4 and and CCONJ fcis-12446 75 5 c3 c3 NOUN fcis-12446 75 6 layers layer NOUN fcis-12446 75 7 are be AUX fcis-12446 75 8 responsible responsible ADJ fcis-12446 75 9 for for ADP fcis-12446 75 10 feature feature NOUN fcis-12446 75 11 extraction extraction NOUN fcis-12446 75 12 from from ADP fcis-12446 75 13 images image NOUN fcis-12446 75 14 , , PUNCT fcis-12446 75 15 while while SCONJ fcis-12446 75 16 the the DET fcis-12446 75 17 sppf sppf ADJ fcis-12446 75 18 layer layer NOUN fcis-12446 75 19 is be AUX fcis-12446 75 20 used use VERB fcis-12446 75 21 to to PART fcis-12446 75 22 convert convert VERB fcis-12446 75 23 feature feature NOUN fcis-12446 75 24 maps map NOUN fcis-12446 75 25 of of ADP fcis-12446 75 26 arbitrary arbitrary ADJ fcis-12446 75 27 sizes size NOUN fcis-12446 75 28 into into ADP fcis-12446 75 29 fixed fix VERB fcis-12446 75 30 - - PUNCT fcis-12446 75 31 size size NOUN fcis-12446 75 32 feature feature NOUN fcis-12446 75 33 vectors vector NOUN fcis-12446 75 34 . . PUNCT fcis-12446 76 1 in in ADP fcis-12446 76 2 figure figure NOUN fcis-12446 76 3 2 2 NUM fcis-12446 76 4 , , PUNCT fcis-12446 76 5 the the DET fcis-12446 76 6 red red ADJ fcis-12446 76 7 modules module NOUN fcis-12446 76 8 represent represent VERB fcis-12446 76 9 the the DET fcis-12446 76 10 introduction introduction NOUN fcis-12446 76 11 of of ADP fcis-12446 76 12 a a DET fcis-12446 76 13 ta ta X fcis-12446 76 14 ( ( PUNCT fcis-12446 76 15 triplet triplet NOUN fcis-12446 76 16 attention attention NOUN fcis-12446 76 17 ) ) PUNCT fcis-12446 76 18 module module NOUN fcis-12446 76 19 after after ADP fcis-12446 76 20 each each DET fcis-12446 76 21 c3 c3 NOUN fcis-12446 76 22 layer layer NOUN fcis-12446 76 23 . . PUNCT fcis-12446 77 1 ta ta X fcis-12446 77 2 is be AUX fcis-12446 77 3 an an DET fcis-12446 77 4 attention attention NOUN fcis-12446 77 5 mechanism mechanism NOUN fcis-12446 77 6 that that PRON fcis-12446 77 7 combines combine VERB fcis-12446 77 8 spatial spatial ADJ fcis-12446 77 9 attention attention NOUN fcis-12446 77 10 with with ADP fcis-12446 77 11 channel channel NOUN fcis-12446 77 12 attention attention NOUN fcis-12446 77 13 , , PUNCT fcis-12446 77 14 allowing allow VERB fcis-12446 77 15 the the DET fcis-12446 77 16 feature feature NOUN fcis-12446 77 17 extraction extraction NOUN fcis-12446 77 18 network network NOUN fcis-12446 77 19 to to PART fcis-12446 77 20 selectively selectively ADV fcis-12446 77 21 focus focus VERB fcis-12446 77 22 on on ADP fcis-12446 77 23 target target NOUN fcis-12446 77 24 regions region NOUN fcis-12446 77 25 containing contain VERB fcis-12446 77 26 important important ADJ fcis-12446 77 27 information information NOUN fcis-12446 77 28 while while SCONJ fcis-12446 77 29 suppressing suppress VERB fcis-12446 77 30 other other ADJ fcis-12446 77 31 irrelevant irrelevant ADJ fcis-12446 77 32 information information NOUN fcis-12446 77 33 . . PUNCT fcis-12446 78 1 this this PRON fcis-12446 78 2 reduces reduce VERB fcis-12446 78 3 the the DET fcis-12446 78 4 impact impact NOUN fcis-12446 78 5 of of ADP fcis-12446 78 6 background background NOUN fcis-12446 78 7 information information NOUN fcis-12446 78 8 on on ADP fcis-12446 78 9 detection detection NOUN fcis-12446 78 10 results result NOUN fcis-12446 78 11 , , PUNCT fcis-12446 78 12 thereby thereby ADV fcis-12446 78 13 improving improve VERB fcis-12446 78 14 the the DET fcis-12446 78 15 model model NOUN fcis-12446 78 16 's 's PART fcis-12446 78 17 ability ability NOUN fcis-12446 78 18 to to PART fcis-12446 78 19 detect detect VERB fcis-12446 78 20 small small ADJ fcis-12446 78 21 targets target NOUN fcis-12446 78 22 . . PUNCT fcis-12446 79 1 the the DET fcis-12446 79 2 specific specific ADJ fcis-12446 79 3 architecture architecture NOUN fcis-12446 79 4 of of ADP fcis-12446 79 5 the the DET fcis-12446 79 6 ta ta PROPN fcis-12446 79 7 attention attention NOUN fcis-12446 79 8 mechanism mechanism NOUN fcis-12446 79 9 is be AUX fcis-12446 79 10 shown show VERB fcis-12446 79 11 in in ADP fcis-12446 79 12 figure figure NOUN fcis-12446 79 13 5 5 NUM fcis-12446 79 14 . . PUNCT fcis-12446 80 1 the the DET fcis-12446 80 2 images image NOUN fcis-12446 80 3 in in ADP fcis-12446 80 4 the the DET fcis-12446 80 5 pest pest NOUN fcis-12446 80 6 dataset dataset NOUN fcis-12446 80 7 are be AUX fcis-12446 80 8 three three NUM fcis-12446 80 9 - - PUNCT fcis-12446 80 10 channel channel NOUN fcis-12446 80 11 images image NOUN fcis-12446 80 12 with with ADP fcis-12446 80 13 a a DET fcis-12446 80 14 tensor tensor NOUN fcis-12446 80 15 shape shape NOUN fcis-12446 80 16 of of ADP fcis-12446 80 17 [ [ X fcis-12446 80 18 h h X fcis-12446 80 19 , , PUNCT fcis-12446 80 20 w w PROPN fcis-12446 80 21 , , PUNCT fcis-12446 80 22 c c NOUN fcis-12446 80 23 ] ] X fcis-12446 80 24 , , PUNCT fcis-12446 80 25 and and CCONJ fcis-12446 80 26 they they PRON fcis-12446 80 27 are be AUX fcis-12446 80 28 input input VERB fcis-12446 80 29 into into ADP fcis-12446 80 30 the the DET fcis-12446 80 31 convolutional convolutional ADJ fcis-12446 80 32 neural neural ADJ fcis-12446 80 33 network network NOUN fcis-12446 80 34 for for ADP fcis-12446 80 35 feature feature NOUN fcis-12446 80 36 extraction extraction NOUN fcis-12446 80 37 . . PUNCT fcis-12446 81 1 as as ADP fcis-12446 81 2 a a DET fcis-12446 81 3 result result NOUN fcis-12446 81 4 , , PUNCT fcis-12446 81 5 the the DET fcis-12446 81 6 shape shape NOUN fcis-12446 81 7 of of ADP fcis-12446 81 8 the the DET fcis-12446 81 9 tensor tensor NOUN fcis-12446 81 10 changes change NOUN fcis-12446 81 11 , , PUNCT fcis-12446 81 12 with with ADP fcis-12446 81 13 reduced reduce VERB fcis-12446 81 14 width width NOUN fcis-12446 81 15 and and CCONJ fcis-12446 81 16 height height NOUN fcis-12446 81 17 but but CCONJ fcis-12446 81 18 increased increased ADJ fcis-12446 81 19 channel channel NOUN fcis-12446 81 20 numbers number NOUN fcis-12446 81 21 , , PUNCT fcis-12446 81 22 as as SCONJ fcis-12446 81 23 shown show VERB fcis-12446 81 24 in in ADP fcis-12446 81 25 the the DET fcis-12446 81 26 input input NOUN fcis-12446 81 27 tensor tensor NOUN fcis-12446 81 28 in in ADP fcis-12446 81 29 figure figure NOUN fcis-12446 81 30 3 3 X fcis-12446 81 31 . . PUNCT fcis-12446 82 1 we we PRON fcis-12446 82 2 feed feed VERB fcis-12446 82 3 the the DET fcis-12446 82 4 input input NOUN fcis-12446 82 5 tensor tensor NOUN fcis-12446 82 6 into into ADP fcis-12446 82 7 three three NUM fcis-12446 82 8 parallel parallel ADJ fcis-12446 82 9 branches branch NOUN fcis-12446 82 10 . . PUNCT fcis-12446 83 1 the the DET fcis-12446 83 2 first first ADJ fcis-12446 83 3 branch branch NOUN fcis-12446 83 4 calculates calculate VERB fcis-12446 83 5 weights weight NOUN fcis-12446 83 6 for for ADP fcis-12446 83 7 regions region NOUN fcis-12446 83 8 of of ADP fcis-12446 83 9 the the DET fcis-12446 83 10 tensor tensor NOUN fcis-12446 83 11 through through ADP fcis-12446 83 12 pooling pooling NOUN fcis-12446 83 13 , , PUNCT fcis-12446 83 14 convolution convolution NOUN fcis-12446 83 15 , , PUNCT fcis-12446 83 16 and and CCONJ fcis-12446 83 17 activation activation NOUN fcis-12446 83 18 , , PUNCT fcis-12446 83 19 and and CCONJ fcis-12446 83 20 then then ADV fcis-12446 83 21 multiplies multiply VERB fcis-12446 83 22 these these DET fcis-12446 83 23 weights weight NOUN fcis-12446 83 24 with with ADP fcis-12446 83 25 the the DET fcis-12446 83 26 original original ADJ fcis-12446 83 27 tensor tensor NOUN fcis-12446 83 28 to to PART fcis-12446 83 29 obtain obtain VERB fcis-12446 83 30 the the DET fcis-12446 83 31 tensor tensor NOUN fcis-12446 83 32 within within ADP fcis-12446 83 33 the the DET fcis-12446 83 34 orange orange PROPN fcis-12446 83 35 region region NOUN fcis-12446 83 36 . . PUNCT fcis-12446 84 1 the the DET fcis-12446 84 2 second second ADJ fcis-12446 84 3 branch branch NOUN fcis-12446 84 4 rotates rotate VERB fcis-12446 84 5 the the DET fcis-12446 84 6 input input NOUN fcis-12446 84 7 tensor tensor NOUN fcis-12446 84 8 counterclockwise counterclockwise NOUN fcis-12446 84 9 by by ADP fcis-12446 84 10 90 90 NUM fcis-12446 84 11 degrees degree NOUN fcis-12446 84 12 along along ADP fcis-12446 84 13 the the DET fcis-12446 84 14 axis axis NOUN fcis-12446 84 15 to to PART fcis-12446 84 16 obtain obtain VERB fcis-12446 84 17 a a DET fcis-12446 84 18 tensor tensor NOUN fcis-12446 84 19 in in ADP fcis-12446 84 20 the the DET fcis-12446 84 21 green green ADJ fcis-12446 84 22 region region NOUN fcis-12446 84 23 . . PUNCT fcis-12446 85 1 it it PRON fcis-12446 85 2 calculates calculate VERB fcis-12446 85 3 weights weight NOUN fcis-12446 85 4 in in ADP fcis-12446 85 5 the the DET fcis-12446 85 6 same same ADJ fcis-12446 85 7 way way NOUN fcis-12446 85 8 as as ADP fcis-12446 85 9 the the DET fcis-12446 85 10 first first ADJ fcis-12446 85 11 branch branch NOUN fcis-12446 85 12 , , PUNCT fcis-12446 85 13 applies apply VERB fcis-12446 85 14 them they PRON fcis-12446 85 15 , , PUNCT fcis-12446 85 16 and and CCONJ fcis-12446 85 17 then then ADV fcis-12446 85 18 rotates rotate VERB fcis-12446 85 19 the the DET fcis-12446 85 20 tensor tensor NOUN fcis-12446 85 21 back back ADV fcis-12446 85 22 to to ADP fcis-12446 85 23 its its PRON fcis-12446 85 24 original original ADJ fcis-12446 85 25 shape shape NOUN fcis-12446 85 26 . . PUNCT fcis-12446 86 1 the the DET fcis-12446 86 2 third third ADJ fcis-12446 86 3 branch branch NOUN fcis-12446 86 4 is be AUX fcis-12446 86 5 similar similar ADJ fcis-12446 86 6 to to ADP fcis-12446 86 7 the the DET fcis-12446 86 8 second second ADJ fcis-12446 86 9 branch branch NOUN fcis-12446 86 10 . . PUNCT fcis-12446 87 1 the the DET fcis-12446 87 2 tensors tensor NOUN fcis-12446 87 3 obtained obtain VERB fcis-12446 87 4 from from ADP fcis-12446 87 5 the the DET fcis-12446 87 6 three three NUM fcis-12446 87 7 branches branch NOUN fcis-12446 87 8 are be AUX fcis-12446 87 9 then then ADV fcis-12446 87 10 added add VERB fcis-12446 87 11 and and CCONJ fcis-12446 87 12 averaged average VERB fcis-12446 87 13 to to PART fcis-12446 87 14 obtain obtain VERB fcis-12446 87 15 the the DET fcis-12446 87 16 output output NOUN fcis-12446 87 17 tensor tensor NOUN fcis-12446 87 18 with with ADP fcis-12446 87 19 cross cross ADJ fcis-12446 87 20 - - ADJ fcis-12446 87 21 dimensional dimensional ADJ fcis-12446 87 22 interactions interaction NOUN fcis-12446 87 23 between between ADP fcis-12446 87 24 spatial spatial ADJ fcis-12446 87 25 and and CCONJ fcis-12446 87 26 temporal temporal ADJ fcis-12446 87 27 attention attention NOUN fcis-12446 87 28 . . PUNCT fcis-12446 88 1 c c NOUN fcis-12446 89 1 h h PROPN fcis-12446 89 2 w w PROPN fcis-12446 89 3 input input PROPN fcis-12446 89 4 w w PROPN fcis-12446 89 5 h h NOUN fcis-12446 89 6 w w PROPN fcis-12446 89 7 c c PROPN fcis-12446 89 8 c c PROPN fcis-12446 89 9 h h PROPN fcis-12446 89 10 avg avg PROPN fcis-12446 89 11     SPACE fcis-12446 89 12 pool+ pool+ PART fcis-12446 89 13 max max PROPN fcis-12446 89 14     SPACE fcis-12446 89 15 pool pool NOUN fcis-12446 89 16 conv conv PROPN fcis-12446 89 17 avg avg PROPN fcis-12446 89 18     SPACE fcis-12446 89 19 pool+ pool+ PART fcis-12446 89 20 max max PROPN fcis-12446 89 21     SPACE fcis-12446 89 22 pool pool NOUN fcis-12446 89 23 conv conv PROPN fcis-12446 89 24 avg avg PROPN fcis-12446 89 25     SPACE fcis-12446 89 26 pool+ pool+ PART fcis-12446 89 27 max max PROPN fcis-12446 89 28     SPACE fcis-12446 89 29 pool pool NOUN fcis-12446 89 30 conv conv NOUN fcis-12446 90 1 c c PROPN fcis-12446 90 2 h h PROPN fcis-12446 90 3 w w PROPN fcis-12446 90 4 c c PROPN fcis-12446 90 5 h h PROPN fcis-12446 90 6 w w PROPN fcis-12446 90 7 c c PROPN fcis-12446 90 8 h h PROPN fcis-12446 90 9 w w PROPN fcis-12446 90 10 sigmo sigmo PROPN fcis-12446 90 11 i i PROPN fcis-12446 90 12 d d PROPN fcis-12446 90 13 sigmo sigmo VERB fcis-12446 90 14 i i PROPN fcis-12446 90 15 d d PROPN fcis-12446 90 16 sigmo sigmo VERB fcis-12446 91 1 i i PROPN fcis-12446 92 1 d d PROPN fcis-12446 92 2 c c PROPN fcis-12446 93 1 h h PROPN fcis-12446 94 1 w w PROPN fcis-12446 94 2 avg avg PROPN fcis-12446 94 3 output output PROPN fcis-12446 94 4 figure figure NOUN fcis-12446 94 5 5 5 NUM fcis-12446 94 6 . . PUNCT fcis-12446 95 1 ta ta X fcis-12446 95 2 ( ( PUNCT fcis-12446 95 3 triplet triplet NOUN fcis-12446 95 4 attention attention NOUN fcis-12446 95 5 ) ) PUNCT fcis-12446 95 6 attention attention NOUN fcis-12446 95 7 diagram diagram NOUN fcis-12446 95 8 . . PUNCT fcis-12446 96 1 to to PART fcis-12446 96 2 make make VERB fcis-12446 96 3 the the DET fcis-12446 96 4 information information NOUN fcis-12446 96 5 in in ADP fcis-12446 96 6 the the DET fcis-12446 96 7 feature feature NOUN fcis-12446 96 8 maps map VERB fcis-12446 96 9 more more ADV fcis-12446 96 10 global global ADJ fcis-12446 96 11 in in ADP fcis-12446 96 12 the the DET fcis-12446 96 13 object object NOUN fcis-12446 96 14 detection detection NOUN fcis-12446 96 15 process process NOUN fcis-12446 96 16 , , PUNCT fcis-12446 96 17 it it PRON fcis-12446 96 18 is be AUX fcis-12446 96 19 common common ADJ fcis-12446 96 20 to to PART fcis-12446 96 21 incorporate incorporate VERB fcis-12446 96 22 feature feature NOUN fcis-12446 96 23 fusion fusion NOUN fcis-12446 96 24 modules module NOUN fcis-12446 96 25 for for ADP fcis-12446 96 26 multi multi ADJ fcis-12446 96 27 - - ADJ fcis-12446 96 28 scale scale ADJ fcis-12446 96 29 feature feature NOUN fcis-12446 96 30 fusion fusion NOUN fcis-12446 96 31 . . PUNCT fcis-12446 97 1 the the DET fcis-12446 97 2 feature feature NOUN fcis-12446 97 3 fusion fusion NOUN fcis-12446 97 4 module module NOUN fcis-12446 97 5 in in ADP fcis-12446 97 6 yolov5 yolov5 NOUN fcis-12446 97 7 is be AUX fcis-12446 97 8 a a DET fcis-12446 97 9 network network NOUN fcis-12446 97 10 structure structure NOUN fcis-12446 97 11 that that PRON fcis-12446 97 12 combines combine VERB fcis-12446 97 13 fpn fpn NOUN fcis-12446 97 14 ( ( PUNCT fcis-12446 97 15 feature feature NOUN fcis-12446 97 16 pyramid pyramid NOUN fcis-12446 97 17 network network NOUN fcis-12446 97 18 ) ) PUNCT fcis-12446 97 19 and and CCONJ fcis-12446 97 20 pan pan NOUN fcis-12446 97 21 ( ( PUNCT fcis-12446 97 22 path path NOUN fcis-12446 97 23 aggregation aggregation NOUN fcis-12446 97 24 network network NOUN fcis-12446 97 25 ) ) PUNCT fcis-12446 97 26 . . PUNCT fcis-12446 98 1 three three NUM fcis-12446 98 2 feature feature NOUN fcis-12446 98 3 maps map NOUN fcis-12446 98 4 with with ADP fcis-12446 98 5 scales scale NOUN fcis-12446 98 6 of of ADP fcis-12446 98 7 [ [ X fcis-12446 98 8 h/8 h/8 NOUN fcis-12446 98 9 , , PUNCT fcis-12446 98 10 w/8 w/8 PROPN fcis-12446 98 11 ] ] PUNCT fcis-12446 98 12 , , PUNCT fcis-12446 98 13 [ [ X fcis-12446 98 14 h/16 h/16 ADJ fcis-12446 98 15 , , PUNCT fcis-12446 98 16 w/16 w/16 PROPN fcis-12446 98 17 ] ] PUNCT fcis-12446 98 18 , , PUNCT fcis-12446 98 19 and and CCONJ fcis-12446 98 20 [ [ X fcis-12446 98 21 h/32 h/32 NOUN fcis-12446 98 22 , , PUNCT fcis-12446 98 23 w/32 w/32 NOUN fcis-12446 98 24 ] ] PUNCT fcis-12446 98 25 are be AUX fcis-12446 98 26 obtained obtain VERB fcis-12446 98 27 from from ADP fcis-12446 98 28 the the DET fcis-12446 98 29 second second ADJ fcis-12446 98 30 and and CCONJ fcis-12446 98 31 third third ADJ fcis-12446 98 32 c3 c3 NOUN fcis-12446 98 33 modules module NOUN fcis-12446 98 34 of of ADP fcis-12446 98 35 the the DET fcis-12446 98 36 backbone backbone NOUN fcis-12446 98 37 64 64 NUM fcis-12446 98 38 network network NOUN fcis-12446 98 39 , , PUNCT fcis-12446 98 40 as as ADV fcis-12446 98 41 well well ADV fcis-12446 98 42 as as ADP fcis-12446 98 43 the the DET fcis-12446 98 44 sppf sppf ADJ fcis-12446 98 45 module module NOUN fcis-12446 98 46 . . PUNCT fcis-12446 99 1 these these DET fcis-12446 99 2 feature feature NOUN fcis-12446 99 3 maps map NOUN fcis-12446 99 4 are be AUX fcis-12446 99 5 then then ADV fcis-12446 99 6 fed feed VERB fcis-12446 99 7 into into ADP fcis-12446 99 8 the the DET fcis-12446 99 9 feature feature NOUN fcis-12446 99 10 fusion fusion NOUN fcis-12446 99 11 module module NOUN fcis-12446 99 12 , , PUNCT fcis-12446 99 13 as as SCONJ fcis-12446 99 14 shown show VERB fcis-12446 99 15 in in ADP fcis-12446 99 16 figure figure NOUN fcis-12446 99 17 6 6 NUM fcis-12446 99 18 : : SYM fcis-12446 99 19 76 76 NUM fcis-12446 99 20 * * SYM fcis-12446 99 21 76 76 NUM fcis-12446 99 22 38 38 NUM fcis-12446 99 23 * * SYM fcis-12446 99 24 38 38 NUM fcis-12446 99 25 19 19 NUM fcis-12446 99 26 * * SYM fcis-12446 99 27 19 19 NUM fcis-12446 99 28 19 19 NUM fcis-12446 99 29 * * NUM fcis-12446 99 30 1938 1938 NUM fcis-12446 99 31 * * SYM fcis-12446 99 32 3876 3876 NUM fcis-12446 99 33 * * SYM fcis-12446 99 34 76 76 NUM fcis-12446 99 35 76 76 NUM fcis-12446 99 36 * * SYM fcis-12446 99 37 76 76 NUM fcis-12446 99 38 38 38 NUM fcis-12446 99 39 * * SYM fcis-12446 99 40 38 38 NUM fcis-12446 99 41 19 19 NUM fcis-12446 99 42 * * SYM fcis-12446 99 43 19 19 NUM fcis-12446 99 44 fpn fpn VERB fcis-12446 99 45 pan pan NOUN fcis-12446 99 46 figure figure NOUN fcis-12446 99 47 6 6 NUM fcis-12446 99 48 . . PUNCT fcis-12446 99 49 feature feature NOUN fcis-12446 99 50 fusion fusion NOUN fcis-12446 99 51 module module NOUN fcis-12446 99 52 . . PUNCT fcis-12446 100 1 the the DET fcis-12446 100 2 orange orange PROPN fcis-12446 100 3 arrows arrow NOUN fcis-12446 100 4 represent represent VERB fcis-12446 100 5 upsampling upsample VERB fcis-12446 100 6 operations operation NOUN fcis-12446 100 7 , , PUNCT fcis-12446 100 8 the the DET fcis-12446 100 9 green green ADJ fcis-12446 100 10 arrows arrow NOUN fcis-12446 100 11 represent represent VERB fcis-12446 100 12 downsampling downsample VERB fcis-12446 100 13 operations operation NOUN fcis-12446 100 14 , , PUNCT fcis-12446 100 15 and and CCONJ fcis-12446 100 16 the the DET fcis-12446 100 17 blue blue NOUN fcis-12446 100 18 dashed dash VERB fcis-12446 100 19 arrows arrow NOUN fcis-12446 100 20 represent represent VERB fcis-12446 100 21 the the DET fcis-12446 100 22 feature feature NOUN fcis-12446 100 23 maps map NOUN fcis-12446 100 24 that that PRON fcis-12446 100 25 will will AUX fcis-12446 100 26 be be AUX fcis-12446 100 27 sent send VERB fcis-12446 100 28 to to ADP fcis-12446 100 29 the the DET fcis-12446 100 30 detection detection NOUN fcis-12446 100 31 head head NOUN fcis-12446 100 32 . . PUNCT fcis-12446 101 1 in in ADP fcis-12446 101 2 the the DET fcis-12446 101 3 fpn fpn NOUN fcis-12446 101 4 module module NOUN fcis-12446 101 5 , , PUNCT fcis-12446 101 6 the the DET fcis-12446 101 7 three three NUM fcis-12446 101 8 feature feature NOUN fcis-12446 101 9 maps map NOUN fcis-12446 101 10 obtained obtain VERB fcis-12446 101 11 from from ADP fcis-12446 101 12 the the DET fcis-12446 101 13 backbone backbone NOUN fcis-12446 101 14 network network NOUN fcis-12446 101 15 undergo undergo VERB fcis-12446 101 16 an an DET fcis-12446 101 17 upsampling upsampling NOUN fcis-12446 101 18 and and CCONJ fcis-12446 101 19 concatenation concatenation NOUN fcis-12446 101 20 operation operation NOUN fcis-12446 101 21 , , PUNCT fcis-12446 101 22 starting start VERB fcis-12446 101 23 from from ADP fcis-12446 101 24 the the DET fcis-12446 101 25 top top ADJ fcis-12446 101 26 - - PUNCT fcis-12446 101 27 level level NOUN fcis-12446 101 28 feature feature NOUN fcis-12446 101 29 map map NOUN fcis-12446 101 30 and and CCONJ fcis-12446 101 31 conveying convey VERB fcis-12446 101 32 semantic semantic ADJ fcis-12446 101 33 information information NOUN fcis-12446 101 34 to to ADP fcis-12446 101 35 the the DET fcis-12446 101 36 bottom bottom ADJ fcis-12446 101 37 - - PUNCT fcis-12446 101 38 level level NOUN fcis-12446 101 39 feature feature NOUN fcis-12446 101 40 maps map NOUN fcis-12446 101 41 for for ADP fcis-12446 101 42 multi multi ADJ fcis-12446 101 43 - - ADJ fcis-12446 101 44 scale scale ADJ fcis-12446 101 45 feature feature NOUN fcis-12446 101 46 fusion fusion NOUN fcis-12446 101 47 . . PUNCT fcis-12446 102 1 pan pan PROPN fcis-12446 102 2 , , PUNCT fcis-12446 102 3 on on ADP fcis-12446 102 4 the the DET fcis-12446 102 5 other other ADJ fcis-12446 102 6 hand hand NOUN fcis-12446 102 7 , , PUNCT fcis-12446 102 8 performs perform VERB fcis-12446 102 9 downsampling downsample VERB fcis-12446 102 10 and and CCONJ fcis-12446 102 11 concatenation concatenation NOUN fcis-12446 102 12 operations operation NOUN fcis-12446 102 13 to to PART fcis-12446 102 14 transfer transfer VERB fcis-12446 102 15 positional positional ADJ fcis-12446 102 16 information information NOUN fcis-12446 102 17 from from ADP fcis-12446 102 18 the the DET fcis-12446 102 19 bottom bottom ADJ fcis-12446 102 20 - - PUNCT fcis-12446 102 21 level level NOUN fcis-12446 102 22 feature feature NOUN fcis-12446 102 23 maps map NOUN fcis-12446 102 24 to to ADP fcis-12446 102 25 the the DET fcis-12446 102 26 top top ADJ fcis-12446 102 27 - - PUNCT fcis-12446 102 28 level level NOUN fcis-12446 102 29 feature feature NOUN fcis-12446 102 30 maps map NOUN fcis-12446 102 31 , , PUNCT fcis-12446 102 32 enhancing enhance VERB fcis-12446 102 33 precision precision NOUN fcis-12446 102 34 in in ADP fcis-12446 102 35 object object NOUN fcis-12446 102 36 localization localization NOUN fcis-12446 102 37 . . PUNCT fcis-12446 103 1 the the DET fcis-12446 103 2 detection detection NOUN fcis-12446 103 3 head head NOUN fcis-12446 103 4 is be AUX fcis-12446 103 5 the the DET fcis-12446 103 6 module module NOUN fcis-12446 103 7 responsible responsible ADJ fcis-12446 103 8 for for ADP fcis-12446 103 9 classifying classify VERB fcis-12446 103 10 and and CCONJ fcis-12446 103 11 localizing localize VERB fcis-12446 103 12 the the DET fcis-12446 103 13 objects object NOUN fcis-12446 103 14 based base VERB fcis-12446 103 15 on on ADP fcis-12446 103 16 the the DET fcis-12446 103 17 feature feature NOUN fcis-12446 103 18 map map NOUN fcis-12446 103 19 information information NOUN fcis-12446 103 20 . . PUNCT fcis-12446 104 1 in in ADP fcis-12446 104 2 the the DET fcis-12446 104 3 traditional traditional ADJ fcis-12446 104 4 yolov5 yolov5 NOUN fcis-12446 104 5 , , PUNCT fcis-12446 104 6 the the DET fcis-12446 104 7 detection detection NOUN fcis-12446 104 8 head head NOUN fcis-12446 104 9 consists consist VERB fcis-12446 104 10 of of ADP fcis-12446 104 11 three three NUM fcis-12446 104 12 parts part NOUN fcis-12446 104 13 , , PUNCT fcis-12446 104 14 each each PRON fcis-12446 104 15 searching search VERB fcis-12446 104 16 for for ADP fcis-12446 104 17 targets target NOUN fcis-12446 104 18 in in ADP fcis-12446 104 19 feature feature NOUN fcis-12446 104 20 maps map NOUN fcis-12446 104 21 of of ADP fcis-12446 104 22 three three NUM fcis-12446 104 23 different different ADJ fcis-12446 104 24 scales scale NOUN fcis-12446 104 25 . . PUNCT fcis-12446 105 1 we we PRON fcis-12446 105 2 have have AUX fcis-12446 105 3 added add VERB fcis-12446 105 4 an an DET fcis-12446 105 5 additional additional ADJ fcis-12446 105 6 detection detection NOUN fcis-12446 105 7 head head NOUN fcis-12446 105 8 , , PUNCT fcis-12446 105 9 making make VERB fcis-12446 105 10 the the DET fcis-12446 105 11 first first ADJ fcis-12446 105 12 detection detection NOUN fcis-12446 105 13 head head NOUN fcis-12446 105 14 responsible responsible ADJ fcis-12446 105 15 for for ADP fcis-12446 105 16 locating locate VERB fcis-12446 105 17 pests pest NOUN fcis-12446 105 18 in in ADP fcis-12446 105 19 densely densely ADV fcis-12446 105 20 populated populated ADJ fcis-12446 105 21 areas area NOUN fcis-12446 105 22 , , PUNCT fcis-12446 105 23 and and CCONJ fcis-12446 105 24 the the DET fcis-12446 105 25 second second ADJ fcis-12446 105 26 detection detection NOUN fcis-12446 105 27 head head NOUN fcis-12446 105 28 for for ADP fcis-12446 105 29 detecting detect VERB fcis-12446 105 30 pests pest NOUN fcis-12446 105 31 within within ADP fcis-12446 105 32 those those DET fcis-12446 105 33 dense dense ADJ fcis-12446 105 34 areas area NOUN fcis-12446 105 35 . . PUNCT fcis-12446 106 1 the the DET fcis-12446 106 2 specific specific ADJ fcis-12446 106 3 details detail NOUN fcis-12446 106 4 are be AUX fcis-12446 106 5 shown show VERB fcis-12446 106 6 in in ADP fcis-12446 106 7 figure figure NOUN fcis-12446 106 8 7 7 NUM fcis-12446 106 9 : : PUNCT fcis-12446 106 10 h1 h1 PROPN fcis-12446 106 11 h2 h2 PROPN fcis-12446 106 12 deconv1deconv2 deconv1deconv2 NOUN fcis-12446 106 13 figure figure NOUN fcis-12446 106 14 7 7 NUM fcis-12446 106 15 . . PUNCT fcis-12446 106 16 detection detection NOUN fcis-12446 106 17 head head NOUN fcis-12446 106 18 module module NOUN fcis-12446 106 19 the the DET fcis-12446 106 20 detection detection NOUN fcis-12446 106 21 head head NOUN fcis-12446 106 22 receives receive VERB fcis-12446 106 23 three three NUM fcis-12446 106 24 feature feature NOUN fcis-12446 106 25 maps map NOUN fcis-12446 106 26 of of ADP fcis-12446 106 27 different different ADJ fcis-12446 106 28 scales scale NOUN fcis-12446 106 29 obtained obtain VERB fcis-12446 106 30 from from ADP fcis-12446 106 31 the the DET fcis-12446 106 32 convolutional convolutional ADJ fcis-12446 106 33 neural neural ADJ fcis-12446 106 34 network network NOUN fcis-12446 106 35 and and CCONJ fcis-12446 106 36 detects detect NOUN fcis-12446 106 37 object object VERB fcis-12446 106 38 targets target NOUN fcis-12446 106 39 based base VERB fcis-12446 106 40 on on ADP fcis-12446 106 41 pre pre ADJ fcis-12446 106 42 - - ADJ fcis-12446 106 43 set set ADJ fcis-12446 106 44 anchor anchor NOUN fcis-12446 106 45 boxes box NOUN fcis-12446 106 46 . . PUNCT fcis-12446 107 1 the the DET fcis-12446 107 2 light light ADJ fcis-12446 107 3 blue blue ADJ fcis-12446 107 4 feature feature NOUN fcis-12446 107 5 map map NOUN fcis-12446 107 6 is be AUX fcis-12446 107 7 used use VERB fcis-12446 107 8 for for ADP fcis-12446 107 9 detecting detect VERB fcis-12446 107 10 small small ADJ fcis-12446 107 11 targets target NOUN fcis-12446 107 12 , , PUNCT fcis-12446 107 13 the the DET fcis-12446 107 14 light light NOUN fcis-12446 107 15 yellow yellow VERB fcis-12446 107 16 one one NUM fcis-12446 107 17 for for ADP fcis-12446 107 18 medium medium ADJ fcis-12446 107 19 - - PUNCT fcis-12446 107 20 sized sized ADJ fcis-12446 107 21 targets target NOUN fcis-12446 107 22 , , PUNCT fcis-12446 107 23 and and CCONJ fcis-12446 107 24 the the DET fcis-12446 107 25 red red ADJ fcis-12446 107 26 small small ADJ fcis-12446 107 27 feature feature NOUN fcis-12446 107 28 map map NOUN fcis-12446 107 29 for for ADP fcis-12446 107 30 detecting detect VERB fcis-12446 107 31 large large ADJ fcis-12446 107 32 targets target NOUN fcis-12446 107 33 . . PUNCT fcis-12446 108 1 therefore therefore ADV fcis-12446 108 2 , , PUNCT fcis-12446 108 3 we we PRON fcis-12446 108 4 obtain obtain VERB fcis-12446 108 5 pest pest NOUN fcis-12446 108 6 - - PUNCT fcis-12446 108 7 dense dense ADJ fcis-12446 108 8 regions region NOUN fcis-12446 108 9 on on ADP fcis-12446 108 10 three three NUM fcis-12446 108 11 feature feature NOUN fcis-12446 108 12 maps map NOUN fcis-12446 108 13 and and CCONJ fcis-12446 108 14 then then ADV fcis-12446 108 15 fuse fuse VERB fcis-12446 108 16 the the DET fcis-12446 108 17 pest pest NOUN fcis-12446 108 18 distribution distribution NOUN fcis-12446 108 19 information information NOUN fcis-12446 108 20 from from ADP fcis-12446 108 21 the the DET fcis-12446 108 22 three three NUM fcis-12446 108 23 feature feature NOUN fcis-12446 108 24 maps map NOUN fcis-12446 108 25 , , PUNCT fcis-12446 108 26 65 65 NUM fcis-12446 108 27 which which PRON fcis-12446 108 28 is be AUX fcis-12446 108 29 sent send VERB fcis-12446 108 30 to to ADP fcis-12446 108 31 the the DET fcis-12446 108 32 detection detection NOUN fcis-12446 108 33 head head NOUN fcis-12446 108 34 for for ADP fcis-12446 108 35 detailed detailed ADJ fcis-12446 108 36 detection detection NOUN fcis-12446 108 37 . . PUNCT fcis-12446 109 1 specifically specifically ADV fcis-12446 109 2 , , PUNCT fcis-12446 109 3 we we PRON fcis-12446 109 4 first first ADV fcis-12446 109 5 perform perform VERB fcis-12446 109 6 deconvolution deconvolution NOUN fcis-12446 109 7 on on ADP fcis-12446 109 8 the the DET fcis-12446 109 9 smallestscale smallestscale ADJ fcis-12446 109 10 feature feature NOUN fcis-12446 109 11 map map NOUN fcis-12446 109 12 and and CCONJ fcis-12446 109 13 add add VERB fcis-12446 109 14 it it PRON fcis-12446 109 15 to to ADP fcis-12446 109 16 the the DET fcis-12446 109 17 corresponding correspond VERB fcis-12446 109 18 positions position NOUN fcis-12446 109 19 of of ADP fcis-12446 109 20 the the DET fcis-12446 109 21 medium medium ADJ fcis-12446 109 22 - - PUNCT fcis-12446 109 23 scale scale NOUN fcis-12446 109 24 feature feature NOUN fcis-12446 109 25 map map NOUN fcis-12446 109 26 . . PUNCT fcis-12446 110 1 we we PRON fcis-12446 110 2 then then ADV fcis-12446 110 3 perform perform VERB fcis-12446 110 4 upsampling upsample VERB fcis-12446 110 5 on on ADP fcis-12446 110 6 the the DET fcis-12446 110 7 obtained obtain VERB fcis-12446 110 8 feature feature NOUN fcis-12446 110 9 map map NOUN fcis-12446 110 10 and and CCONJ fcis-12446 110 11 add add VERB fcis-12446 110 12 it it PRON fcis-12446 110 13 to to ADP fcis-12446 110 14 the the DET fcis-12446 110 15 corresponding correspond VERB fcis-12446 110 16 positions position NOUN fcis-12446 110 17 of of ADP fcis-12446 110 18 the the DET fcis-12446 110 19 large large ADJ fcis-12446 110 20 - - PUNCT fcis-12446 110 21 scale scale NOUN fcis-12446 110 22 feature feature NOUN fcis-12446 110 23 map map NOUN fcis-12446 110 24 to to PART fcis-12446 110 25 obtain obtain VERB fcis-12446 110 26 the the DET fcis-12446 110 27 final final ADJ fcis-12446 110 28 feature feature NOUN fcis-12446 110 29 map map NOUN fcis-12446 110 30 after after ADP fcis-12446 110 31 multi multi ADJ fcis-12446 110 32 - - ADJ fcis-12446 110 33 scale scale ADJ fcis-12446 110 34 fusion fusion NOUN fcis-12446 110 35 . . PUNCT fcis-12446 111 1 this this DET fcis-12446 111 2 process process NOUN fcis-12446 111 3 fuses fuse VERB fcis-12446 111 4 the the DET fcis-12446 111 5 target target NOUN fcis-12446 111 6 information information NOUN fcis-12446 111 7 carried carry VERB fcis-12446 111 8 by by ADP fcis-12446 111 9 the the DET fcis-12446 111 10 three three NUM fcis-12446 111 11 different different ADJ fcis-12446 111 12 - - PUNCT fcis-12446 111 13 scale scale NOUN fcis-12446 111 14 feature feature NOUN fcis-12446 111 15 maps map NOUN fcis-12446 111 16 and and CCONJ fcis-12446 111 17 yields yield VERB fcis-12446 111 18 the the DET fcis-12446 111 19 dense dense ADJ fcis-12446 111 20 pest pest NOUN fcis-12446 111 21 distribution distribution NOUN fcis-12446 111 22 areas area NOUN fcis-12446 111 23 , , PUNCT fcis-12446 111 24 which which PRON fcis-12446 111 25 are be AUX fcis-12446 111 26 then then ADV fcis-12446 111 27 sent send VERB fcis-12446 111 28 to to ADP fcis-12446 111 29 the the DET fcis-12446 111 30 detection detection NOUN fcis-12446 111 31 head head NOUN fcis-12446 111 32 , , PUNCT fcis-12446 111 33 achieving achieve VERB fcis-12446 111 34 detection detection NOUN fcis-12446 111 35 that that PRON fcis-12446 111 36 focuses focus VERB fcis-12446 111 37 only only ADV fcis-12446 111 38 on on ADP fcis-12446 111 39 dense dense ADJ fcis-12446 111 40 areas area NOUN fcis-12446 111 41 and and CCONJ fcis-12446 111 42 outliers outlier NOUN fcis-12446 111 43 . . PUNCT fcis-12446 112 1 with with ADP fcis-12446 112 2 the the DET fcis-12446 112 3 development development NOUN fcis-12446 112 4 of of ADP fcis-12446 112 5 multi multi ADJ fcis-12446 112 6 - - ADJ fcis-12446 112 7 task task ADJ fcis-12446 112 8 loss loss NOUN fcis-12446 112 9 functions function NOUN fcis-12446 112 10 , , PUNCT fcis-12446 112 11 this this DET fcis-12446 112 12 method method NOUN fcis-12446 112 13 becomes become VERB fcis-12446 112 14 more more ADV fcis-12446 112 15 practical practical ADJ fcis-12446 112 16 , , PUNCT fcis-12446 112 17 and and CCONJ fcis-12446 112 18 we we PRON fcis-12446 112 19 adjust adjust VERB fcis-12446 112 20 the the DET fcis-12446 112 21 weights weight NOUN fcis-12446 112 22 between between ADP fcis-12446 112 23 the the DET fcis-12446 112 24 detection detection NOUN fcis-12446 112 25 head head NOUN fcis-12446 112 26 and and CCONJ fcis-12446 112 27 the the DET fcis-12446 112 28 use use NOUN fcis-12446 112 29 of of ADP fcis-12446 112 30 different different ADJ fcis-12446 112 31 loss loss NOUN fcis-12446 112 32 functions function NOUN fcis-12446 112 33 to to PART fcis-12446 112 34 achieve achieve VERB fcis-12446 112 35 better well ADJ fcis-12446 112 36 results result NOUN fcis-12446 112 37 . . PUNCT fcis-12446 113 1 to to PART fcis-12446 113 2 improve improve VERB fcis-12446 113 3 convergence convergence NOUN fcis-12446 113 4 speed speed NOUN fcis-12446 113 5 and and CCONJ fcis-12446 113 6 final final ADJ fcis-12446 113 7 accuracy accuracy NOUN fcis-12446 113 8 , , PUNCT fcis-12446 113 9 we we PRON fcis-12446 113 10 calculate calculate VERB fcis-12446 113 11 the the DET fcis-12446 113 12 most most ADV fcis-12446 113 13 suitable suitable ADJ fcis-12446 113 14 anchor anchor NOUN fcis-12446 113 15 box box NOUN fcis-12446 113 16 values value NOUN fcis-12446 113 17 for for ADP fcis-12446 113 18 the the DET fcis-12446 113 19 real real ADJ fcis-12446 113 20 boxes box NOUN fcis-12446 113 21 in in ADP fcis-12446 113 22 the the DET fcis-12446 113 23 dataset dataset NOUN fcis-12446 113 24 and and CCONJ fcis-12446 113 25 the the DET fcis-12446 113 26 dense dense ADJ fcis-12446 113 27 distribution distribution NOUN fcis-12446 113 28 real real ADJ fcis-12446 113 29 boxes box NOUN fcis-12446 113 30 through through ADP fcis-12446 113 31 clustering clustering NOUN fcis-12446 113 32 . . PUNCT fcis-12446 114 1 for for ADP fcis-12446 114 2 each each DET fcis-12446 114 3 feature feature NOUN fcis-12446 114 4 map map NOUN fcis-12446 114 5 , , PUNCT fcis-12446 114 6 we we PRON fcis-12446 114 7 assign assign VERB fcis-12446 114 8 three three NUM fcis-12446 114 9 anchor anchor NOUN fcis-12446 114 10 boxes box NOUN fcis-12446 114 11 with with ADP fcis-12446 114 12 different different ADJ fcis-12446 114 13 aspect aspect NOUN fcis-12446 114 14 ratios ratio NOUN fcis-12446 114 15 , , PUNCT fcis-12446 114 16 with with ADP fcis-12446 114 17 ratios ratio NOUN fcis-12446 114 18 of of ADP fcis-12446 114 19 [ [ X fcis-12446 114 20 116 116 NUM fcis-12446 114 21 , , PUNCT fcis-12446 114 22 90 90 NUM fcis-12446 114 23 ] ] PUNCT fcis-12446 114 24 , , PUNCT fcis-12446 114 25 [ [ X fcis-12446 114 26 156 156 NUM fcis-12446 114 27 , , PUNCT fcis-12446 114 28 198 198 NUM fcis-12446 114 29 ] ] PUNCT fcis-12446 114 30 , , PUNCT fcis-12446 114 31 and and CCONJ fcis-12446 114 32 [ [ X fcis-12446 114 33 373 373 NUM fcis-12446 114 34 , , PUNCT fcis-12446 114 35 326 326 NUM fcis-12446 114 36 ] ] PUNCT fcis-12446 114 37 . . PUNCT fcis-12446 115 1 the the DET fcis-12446 115 2 detection detection NOUN fcis-12446 115 3 head head NOUN fcis-12446 115 4 is be AUX fcis-12446 115 5 responsible responsible ADJ fcis-12446 115 6 for for ADP fcis-12446 115 7 classifying classify VERB fcis-12446 115 8 and and CCONJ fcis-12446 115 9 regressing regress VERB fcis-12446 115 10 on on ADP fcis-12446 115 11 one one NUM fcis-12446 115 12 scale scale NOUN fcis-12446 115 13 of of ADP fcis-12446 115 14 feature feature NOUN fcis-12446 115 15 map map NOUN fcis-12446 115 16 , , PUNCT fcis-12446 115 17 and and CCONJ fcis-12446 115 18 we we PRON fcis-12446 115 19 have have AUX fcis-12446 115 20 preset preset VERB fcis-12446 115 21 nine nine NUM fcis-12446 115 22 anchor anchor NOUN fcis-12446 115 23 boxes box NOUN fcis-12446 115 24 on on ADP fcis-12446 115 25 it it PRON fcis-12446 115 26 , , PUNCT fcis-12446 115 27 with with ADP fcis-12446 115 28 ratios ratio NOUN fcis-12446 115 29 of of ADP fcis-12446 115 30 [ [ X fcis-12446 115 31 5 5 NUM fcis-12446 115 32 , , PUNCT fcis-12446 115 33 6 6 NUM fcis-12446 115 34 , , PUNCT fcis-12446 115 35 8 8 NUM fcis-12446 115 36 , , PUNCT fcis-12446 115 37 14 14 NUM fcis-12446 115 38 , , PUNCT fcis-12446 115 39 15 15 NUM fcis-12446 115 40 , , PUNCT fcis-12446 115 41 11 11 NUM fcis-12446 115 42 ] ] PUNCT fcis-12446 115 43 , , PUNCT fcis-12446 115 44 [ [ X fcis-12446 115 45 10 10 NUM fcis-12446 115 46 , , PUNCT fcis-12446 115 47 13 13 NUM fcis-12446 115 48 , , PUNCT fcis-12446 115 49 16 16 NUM fcis-12446 115 50 , , PUNCT fcis-12446 115 51 30 30 NUM fcis-12446 115 52 , , PUNCT fcis-12446 115 53 33 33 NUM fcis-12446 115 54 , , PUNCT fcis-12446 115 55 23 23 NUM fcis-12446 115 56 ] ] PUNCT fcis-12446 115 57 , , PUNCT fcis-12446 115 58 and and CCONJ fcis-12446 115 59 [ [ X fcis-12446 115 60 30 30 NUM fcis-12446 115 61 , , PUNCT fcis-12446 115 62 61 61 NUM fcis-12446 115 63 , , PUNCT fcis-12446 115 64 62 62 NUM fcis-12446 115 65 , , PUNCT fcis-12446 115 66 45 45 NUM fcis-12446 115 67 , , PUNCT fcis-12446 115 68 59 59 NUM fcis-12446 115 69 , , PUNCT fcis-12446 115 70 119 119 NUM fcis-12446 115 71 ] ] PUNCT fcis-12446 115 72 . . PUNCT fcis-12446 116 1 since since SCONJ fcis-12446 116 2 the the DET fcis-12446 116 3 annotated annotate VERB fcis-12446 116 4 information information NOUN fcis-12446 116 5 for for ADP fcis-12446 116 6 dense dense ADJ fcis-12446 116 7 distribution distribution NOUN fcis-12446 116 8 targets target NOUN fcis-12446 116 9 includes include VERB fcis-12446 116 10 both both CCONJ fcis-12446 116 11 well well ADV fcis-12446 116 12 - - PUNCT fcis-12446 116 13 separated separate VERB fcis-12446 116 14 pest pest NOUN fcis-12446 116 15 targets target NOUN fcis-12446 116 16 with with ADP fcis-12446 116 17 a a DET fcis-12446 116 18 small small ADJ fcis-12446 116 19 proportion proportion NOUN fcis-12446 116 20 in in ADP fcis-12446 116 21 the the DET fcis-12446 116 22 original original ADJ fcis-12446 116 23 image image NOUN fcis-12446 116 24 and and CCONJ fcis-12446 116 25 large large ADJ fcis-12446 116 26 and and CCONJ fcis-12446 116 27 medium medium ADJ fcis-12446 116 28 - - PUNCT fcis-12446 116 29 sized sized ADJ fcis-12446 116 30 true true ADJ fcis-12446 116 31 target target NOUN fcis-12446 116 32 boxes box NOUN fcis-12446 116 33 distributed distribute VERB fcis-12446 116 34 in in ADP fcis-12446 116 35 clusters cluster NOUN fcis-12446 116 36 , , PUNCT fcis-12446 116 37 using use VERB fcis-12446 116 38 the the DET fcis-12446 116 39 default default NOUN fcis-12446 116 40 anchor anchor NOUN fcis-12446 116 41 boxes box NOUN fcis-12446 116 42 in in ADP fcis-12446 116 43 yolov5 yolov5 NOUN fcis-12446 116 44 for for ADP fcis-12446 116 45 dense dense ADJ fcis-12446 116 46 area area NOUN fcis-12446 116 47 detection detection NOUN fcis-12446 116 48 is be AUX fcis-12446 116 49 sufficient sufficient ADJ fcis-12446 116 50 to to PART fcis-12446 116 51 produce produce VERB fcis-12446 116 52 good good ADJ fcis-12446 116 53 detection detection NOUN fcis-12446 116 54 results result NOUN fcis-12446 116 55 . . PUNCT fcis-12446 117 1 in in ADP fcis-12446 117 2 the the DET fcis-12446 117 3 only only ADJ fcis-12446 117 4 feature feature NOUN fcis-12446 117 5 map map NOUN fcis-12446 117 6 faced face VERB fcis-12446 117 7 by by ADP fcis-12446 117 8 the the DET fcis-12446 117 9 detection detection NOUN fcis-12446 117 10 head head NOUN fcis-12446 117 11 , , PUNCT fcis-12446 117 12 pests pest NOUN fcis-12446 117 13 are be AUX fcis-12446 117 14 concentrated concentrate VERB fcis-12446 117 15 , , PUNCT fcis-12446 117 16 and and CCONJ fcis-12446 117 17 using use VERB fcis-12446 117 18 only only ADV fcis-12446 117 19 three three NUM fcis-12446 117 20 anchor anchor NOUN fcis-12446 117 21 boxes box NOUN fcis-12446 117 22 may may AUX fcis-12446 117 23 lead lead VERB fcis-12446 117 24 to to ADP fcis-12446 117 25 a a DET fcis-12446 117 26 low low ADJ fcis-12446 117 27 recall recall NOUN fcis-12446 117 28 rate rate NOUN fcis-12446 117 29 for for ADP fcis-12446 117 30 small small ADJ fcis-12446 117 31 target target NOUN fcis-12446 117 32 detection detection NOUN fcis-12446 117 33 . . PUNCT fcis-12446 118 1 therefore therefore ADV fcis-12446 118 2 , , PUNCT fcis-12446 118 3 we we PRON fcis-12446 118 4 need need VERB fcis-12446 118 5 to to PART fcis-12446 118 6 lay lay VERB fcis-12446 118 7 denser dense ADJ fcis-12446 118 8 anchor anchor NOUN fcis-12446 118 9 boxes box NOUN fcis-12446 118 10 . . PUNCT fcis-12446 119 1 however however ADV fcis-12446 119 2 , , PUNCT fcis-12446 119 3 this this DET fcis-12446 119 4 operation operation NOUN fcis-12446 119 5 introduces introduce VERB fcis-12446 119 6 more more ADJ fcis-12446 119 7 parameters parameter NOUN fcis-12446 119 8 and and CCONJ fcis-12446 119 9 computational computational ADJ fcis-12446 119 10 complexity complexity NOUN fcis-12446 119 11 . . PUNCT fcis-12446 120 1 to to PART fcis-12446 120 2 address address VERB fcis-12446 120 3 this this DET fcis-12446 120 4 issue issue NOUN fcis-12446 120 5 , , PUNCT fcis-12446 120 6 we we PRON fcis-12446 120 7 made make VERB fcis-12446 120 8 some some DET fcis-12446 120 9 improvements improvement NOUN fcis-12446 120 10 in in ADP fcis-12446 120 11 the the DET fcis-12446 120 12 positive positive ADJ fcis-12446 120 13 and and CCONJ fcis-12446 120 14 negative negative ADJ fcis-12446 120 15 sample sample NOUN fcis-12446 120 16 matching matching NOUN fcis-12446 120 17 process process NOUN fcis-12446 120 18 . . PUNCT fcis-12446 121 1 the the DET fcis-12446 121 2 positive positive ADJ fcis-12446 121 3 and and CCONJ fcis-12446 121 4 negative negative ADJ fcis-12446 121 5 sample sample NOUN fcis-12446 121 6 selection selection NOUN fcis-12446 121 7 scheme scheme NOUN fcis-12446 121 8 for for ADP fcis-12446 121 9 the the DET fcis-12446 121 10 detection detection NOUN fcis-12446 121 11 head head NOUN fcis-12446 121 12 is be AUX fcis-12446 121 13 consistent consistent ADJ fcis-12446 121 14 with with ADP fcis-12446 121 15 yolov5 yolov5 NOUN fcis-12446 121 16 , , PUNCT fcis-12446 121 17 requiring require VERB fcis-12446 121 18 that that SCONJ fcis-12446 121 19 the the DET fcis-12446 121 20 overlap overlap NOUN fcis-12446 121 21 between between ADP fcis-12446 121 22 anchor anchor NOUN fcis-12446 121 23 boxes box NOUN fcis-12446 121 24 and and CCONJ fcis-12446 121 25 real real ADJ fcis-12446 121 26 boxes box NOUN fcis-12446 121 27 is be AUX fcis-12446 121 28 greater great ADJ fcis-12446 121 29 than than ADP fcis-12446 121 30 a a DET fcis-12446 121 31 threshold threshold NOUN fcis-12446 121 32 and and CCONJ fcis-12446 121 33 the the DET fcis-12446 121 34 aspect aspect NOUN fcis-12446 121 35 ratios ratio NOUN fcis-12446 121 36 of of ADP fcis-12446 121 37 the the DET fcis-12446 121 38 two two NUM fcis-12446 121 39 are be AUX fcis-12446 121 40 not not PART fcis-12446 121 41 too too ADV fcis-12446 121 42 different different ADJ fcis-12446 121 43 . . PUNCT fcis-12446 122 1 for for ADP fcis-12446 122 2 the the DET fcis-12446 122 3 detection detection NOUN fcis-12446 122 4 head head NOUN fcis-12446 122 5 , , PUNCT fcis-12446 122 6 we we PRON fcis-12446 122 7 propose propose VERB fcis-12446 122 8 a a DET fcis-12446 122 9 new new ADJ fcis-12446 122 10 matching matching NOUN fcis-12446 122 11 requirement requirement NOUN fcis-12446 122 12 : : PUNCT fcis-12446 122 13 we we PRON fcis-12446 122 14 first first ADV fcis-12446 122 15 determine determine VERB fcis-12446 122 16 whether whether SCONJ fcis-12446 122 17 the the DET fcis-12446 122 18 center center NOUN fcis-12446 122 19 point point NOUN fcis-12446 122 20 of of ADP fcis-12446 122 21 the the DET fcis-12446 122 22 pre pre ADJ fcis-12446 122 23 - - ADJ fcis-12446 122 24 set set ADJ fcis-12446 122 25 anchor anchor NOUN fcis-12446 122 26 box box NOUN fcis-12446 122 27 falls fall VERB fcis-12446 122 28 within within ADP fcis-12446 122 29 the the DET fcis-12446 122 30 dense dense ADJ fcis-12446 122 31 pest pest NOUN fcis-12446 122 32 area area NOUN fcis-12446 122 33 output output NOUN fcis-12446 122 34 by by ADP fcis-12446 122 35 the the DET fcis-12446 122 36 detection detection NOUN fcis-12446 122 37 head head NOUN fcis-12446 122 38 . . PUNCT fcis-12446 123 1 when when SCONJ fcis-12446 123 2 it it PRON fcis-12446 123 3 falls fall VERB fcis-12446 123 4 within within ADP fcis-12446 123 5 the the DET fcis-12446 123 6 dense dense ADJ fcis-12446 123 7 area area NOUN fcis-12446 123 8 and and CCONJ fcis-12446 123 9 meets meet VERB fcis-12446 123 10 the the DET fcis-12446 123 11 two two NUM fcis-12446 123 12 conditions condition NOUN fcis-12446 123 13 for for ADP fcis-12446 123 14 matching match VERB fcis-12446 123 15 positive positive ADJ fcis-12446 123 16 samples sample NOUN fcis-12446 123 17 of of ADP fcis-12446 123 18 the the DET fcis-12446 123 19 detection detection NOUN fcis-12446 123 20 head head NOUN fcis-12446 123 21 , , PUNCT fcis-12446 123 22 it it PRON fcis-12446 123 23 is be AUX fcis-12446 123 24 considered consider VERB fcis-12446 123 25 a a DET fcis-12446 123 26 positive positive ADJ fcis-12446 123 27 sample sample NOUN fcis-12446 123 28 and and CCONJ fcis-12446 123 29 the the DET fcis-12446 123 30 loss loss NOUN fcis-12446 123 31 function function NOUN fcis-12446 123 32 is be AUX fcis-12446 123 33 calculated calculate VERB fcis-12446 123 34 ; ; PUNCT fcis-12446 123 35 otherwise otherwise ADV fcis-12446 123 36 , , PUNCT fcis-12446 123 37 it it PRON fcis-12446 123 38 will will AUX fcis-12446 123 39 not not PART fcis-12446 123 40 be be AUX fcis-12446 123 41 matched match VERB fcis-12446 123 42 as as ADP fcis-12446 123 43 a a DET fcis-12446 123 44 positive positive ADJ fcis-12446 123 45 or or CCONJ fcis-12446 123 46 negative negative ADJ fcis-12446 123 47 sample sample NOUN fcis-12446 123 48 . . PUNCT fcis-12446 124 1 therefore therefore ADV fcis-12446 124 2 , , PUNCT fcis-12446 124 3 by by ADP fcis-12446 124 4 first first ADV fcis-12446 124 5 filtering filter VERB fcis-12446 124 6 the the DET fcis-12446 124 7 detection detection NOUN fcis-12446 124 8 areas area NOUN fcis-12446 124 9 , , PUNCT fcis-12446 124 10 finding find VERB fcis-12446 124 11 the the DET fcis-12446 124 12 range range NOUN fcis-12446 124 13 with with ADP fcis-12446 124 14 dense dense ADJ fcis-12446 124 15 targets target NOUN fcis-12446 124 16 , , PUNCT fcis-12446 124 17 and and CCONJ fcis-12446 124 18 then then ADV fcis-12446 124 19 using use VERB fcis-12446 124 20 multiple multiple ADJ fcis-12446 124 21 anchor anchor NOUN fcis-12446 124 22 boxes box NOUN fcis-12446 124 23 for for ADP fcis-12446 124 24 multi multi ADJ fcis-12446 124 25 - - ADJ fcis-12446 124 26 target target ADJ fcis-12446 124 27 detection detection NOUN fcis-12446 124 28 , , PUNCT fcis-12446 124 29 we we PRON fcis-12446 124 30 greatly greatly ADV fcis-12446 124 31 increase increase VERB fcis-12446 124 32 the the DET fcis-12446 124 33 recall recall NOUN fcis-12446 124 34 rate rate NOUN fcis-12446 124 35 of of ADP fcis-12446 124 36 dense dense ADJ fcis-12446 124 37 pest pest NOUN fcis-12446 124 38 detection detection NOUN fcis-12446 124 39 at at ADP fcis-12446 124 40 a a DET fcis-12446 124 41 relatively relatively ADV fcis-12446 124 42 low low ADJ fcis-12446 124 43 cost cost NOUN fcis-12446 124 44 . . PUNCT fcis-12446 125 1 4 4 X fcis-12446 125 2 . . X fcis-12446 125 3 experimental experimental ADJ fcis-12446 125 4 and and CCONJ fcis-12446 125 5 dataset dataset NOUN fcis-12446 125 6 we we PRON fcis-12446 125 7 started start VERB fcis-12446 125 8 training train VERB fcis-12446 125 9 a a DET fcis-12446 125 10 pest pest NOUN fcis-12446 125 11 detection detection NOUN fcis-12446 125 12 network network NOUN fcis-12446 125 13 from from ADP fcis-12446 125 14 scratch scratch NOUN fcis-12446 125 15 , , PUNCT fcis-12446 125 16 making make VERB fcis-12446 125 17 appropriate appropriate ADJ fcis-12446 125 18 adjustments adjustment NOUN fcis-12446 125 19 to to ADP fcis-12446 125 20 hyperparameters hyperparameter NOUN fcis-12446 125 21 during during ADP fcis-12446 125 22 the the DET fcis-12446 125 23 process process NOUN fcis-12446 125 24 . . PUNCT fcis-12446 126 1 our our PRON fcis-12446 126 2 experiments experiment NOUN fcis-12446 126 3 were be AUX fcis-12446 126 4 conducted conduct VERB fcis-12446 126 5 using use VERB fcis-12446 126 6 the the DET fcis-12446 126 7 ap ap PROPN fcis-12446 126 8 dataset dataset NOUN fcis-12446 126 9 , , PUNCT fcis-12446 126 10 which which PRON fcis-12446 126 11 consists consist VERB fcis-12446 126 12 of of ADP fcis-12446 126 13 aphid aphid NOUN fcis-12446 126 14 images image NOUN fcis-12446 126 15 collected collect VERB fcis-12446 126 16 in in ADP fcis-12446 126 17 agricultural agricultural ADJ fcis-12446 126 18 environments environment NOUN fcis-12446 126 19 , , PUNCT fcis-12446 126 20 and and CCONJ fcis-12446 126 21 the the DET fcis-12446 126 22 composition composition NOUN fcis-12446 126 23 of of ADP fcis-12446 126 24 the the DET fcis-12446 126 25 dataset dataset NOUN fcis-12446 126 26 's 's PART fcis-12446 126 27 images image NOUN fcis-12446 126 28 is be AUX fcis-12446 126 29 presented present VERB fcis-12446 126 30 in in ADP fcis-12446 126 31 table table NOUN fcis-12446 126 32 1 1 NUM fcis-12446 126 33 . . PUNCT fcis-12446 126 34 table table NOUN fcis-12446 126 35 1 1 NUM fcis-12446 126 36 . . PUNCT fcis-12446 127 1 composition composition NOUN fcis-12446 127 2 of of ADP fcis-12446 127 3 pest pest NOUN fcis-12446 127 4 information information NOUN fcis-12446 127 5 in in ADP fcis-12446 127 6 the the DET fcis-12446 127 7 ap ap PROPN fcis-12446 127 8 dataset dataset PROPN fcis-12446 127 9 class class NOUN fcis-12446 127 10 ratito ratito ADJ fcis-12446 127 11 number number NOUN fcis-12446 127 12 object object NOUN fcis-12446 127 13 greenbug greenbug NOUN fcis-12446 127 14 8.6 8.6 NUM fcis-12446 127 15 % % NOUN fcis-12446 127 16 426 426 NUM fcis-12446 127 17 26380 26380 NUM fcis-12446 127 18 aphid aphid NOUN fcis-12446 127 19 6.0 6.0 NUM fcis-12446 127 20 % % NOUN fcis-12446 127 21 151 151 NUM fcis-12446 127 22 61050 61050 NUM fcis-12446 127 23 sitobion sitobion NOUN fcis-12446 127 24 avenae avenae NOUN fcis-12446 127 25 4.2 4.2 NUM fcis-12446 127 26 % % NOUN fcis-12446 127 27 34 34 NUM fcis-12446 127 28 22538 22538 NUM fcis-12446 127 29 all all ADV fcis-12446 127 30 -----611 -----611 ADP fcis-12446 127 31 109968 109968 NUM fcis-12446 127 32 the the DET fcis-12446 127 33 method method NOUN fcis-12446 127 34 proposed propose VERB fcis-12446 127 35 in in ADP fcis-12446 127 36 this this DET fcis-12446 127 37 paper paper NOUN fcis-12446 127 38 , , PUNCT fcis-12446 127 39 along along ADP fcis-12446 127 40 with with ADP fcis-12446 127 41 the the DET fcis-12446 127 42 related relate VERB fcis-12446 127 43 comparative comparative ADJ fcis-12446 127 44 experiments experiment NOUN fcis-12446 127 45 , , PUNCT fcis-12446 127 46 was be AUX fcis-12446 127 47 implemented implement VERB fcis-12446 127 48 using use VERB fcis-12446 127 49 the the DET fcis-12446 127 50 pycharm pycharm NOUN fcis-12446 127 51 integrated integrate VERB fcis-12446 127 52 development development NOUN fcis-12446 127 53 environment environment NOUN fcis-12446 127 54 and and CCONJ fcis-12446 127 55 code code NOUN fcis-12446 127 56 written write VERB fcis-12446 127 57 using use VERB fcis-12446 127 58 the the DET fcis-12446 127 59 pytorch pytorch NOUN fcis-12446 127 60 1.10.2 1.10.2 NUM fcis-12446 127 61 framework framework NOUN fcis-12446 127 62 with with ADP fcis-12446 127 63 python python PROPN fcis-12446 127 64 3.6 3.6 NUM fcis-12446 127 65 api api NOUN fcis-12446 127 66 . . PUNCT fcis-12446 128 1 we we PRON fcis-12446 128 2 used use VERB fcis-12446 128 3 an an DET fcis-12446 128 4 18.04 18.04 NUM fcis-12446 128 5 version version NOUN fcis-12446 128 6 of of ADP fcis-12446 128 7 the the DET fcis-12446 128 8 ubuntu ubuntu ADJ fcis-12446 128 9 host host NOUN fcis-12446 128 10 on on ADP fcis-12446 128 11 a a DET fcis-12446 128 12 linux linux PROPN fcis-12446 128 13 server server NOUN fcis-12446 128 14 , , PUNCT fcis-12446 128 15 equipped equip VERB fcis-12446 128 16 with with ADP fcis-12446 128 17 a a DET fcis-12446 128 18 tesla tesla PROPN fcis-12446 128 19 t4 t4 PROPN fcis-12446 128 20 gpu gpu PROPN fcis-12446 128 21 processor processor NOUN fcis-12446 128 22 with with ADP fcis-12446 128 23 15 15 NUM fcis-12446 128 24 gb gb NOUN fcis-12446 128 25 of of ADP fcis-12446 128 26 memory memory NOUN fcis-12446 128 27 and and CCONJ fcis-12446 128 28 48 48 NUM fcis-12446 128 29 cpu cpu NOUN fcis-12446 128 30 processors processor NOUN fcis-12446 128 31 . . PUNCT fcis-12446 129 1 the the DET fcis-12446 129 2 training training NOUN fcis-12446 129 3 was be AUX fcis-12446 129 4 performed perform VERB fcis-12446 129 5 using use VERB fcis-12446 129 6 the the DET fcis-12446 129 7 sgd sgd NOUN fcis-12446 129 8 algorithm algorithm NOUN fcis-12446 129 9 on on ADP fcis-12446 129 10 the the DET fcis-12446 129 11 t4 t4 PROPN fcis-12446 129 12 processor processor NOUN fcis-12446 129 13 , , PUNCT fcis-12446 129 14 running run VERB fcis-12446 129 15 for for ADP fcis-12446 129 16 a a DET fcis-12446 129 17 total total NOUN fcis-12446 129 18 of of ADP fcis-12446 129 19 500 500 NUM fcis-12446 129 20 epochs epoch NOUN fcis-12446 129 21 , , PUNCT fcis-12446 129 22 with with ADP fcis-12446 129 23 each each DET fcis-12446 129 24 batch batch NOUN fcis-12446 129 25 containing contain VERB fcis-12446 129 26 16 16 NUM fcis-12446 129 27 images image NOUN fcis-12446 129 28 . . PUNCT fcis-12446 130 1 the the DET fcis-12446 130 2 initial initial ADJ fcis-12446 130 3 learning learning NOUN fcis-12446 130 4 rate rate NOUN fcis-12446 130 5 was be AUX fcis-12446 130 6 set set VERB fcis-12446 130 7 to to ADP fcis-12446 130 8 a a DET fcis-12446 130 9 certain certain ADJ fcis-12446 130 10 value value NOUN fcis-12446 130 11 and and CCONJ fcis-12446 130 12 then then ADV fcis-12446 130 13 reduced reduce VERB fcis-12446 130 14 by by ADP fcis-12446 130 15 a a DET fcis-12446 130 16 factor factor NOUN fcis-12446 130 17 of of ADP fcis-12446 130 18 10 10 NUM fcis-12446 130 19 every every DET fcis-12446 130 20 1000 1000 NUM fcis-12446 130 21 iterations iteration NOUN fcis-12446 130 22 using use VERB fcis-12446 130 23 learning learn VERB fcis-12446 130 24 rate rate NOUN fcis-12446 130 25 decay decay NOUN fcis-12446 130 26 . . PUNCT fcis-12446 131 1 we we PRON fcis-12446 131 2 incorporated incorporate VERB fcis-12446 131 3 an an DET fcis-12446 131 4 attention attention NOUN fcis-12446 131 5 mechanism mechanism NOUN fcis-12446 131 6 into into ADP fcis-12446 131 7 the the DET fcis-12446 131 8 yolov5 yolov5 NOUN fcis-12446 131 9 backbone backbone NOUN fcis-12446 131 10 network network NOUN fcis-12446 131 11 and and CCONJ fcis-12446 131 12 maintained maintain VERB fcis-12446 131 13 the the DET fcis-12446 131 14 fpn+pan fpn+pan PROPN fcis-12446 131 15 structure structure NOUN fcis-12446 131 16 . . PUNCT fcis-12446 132 1 in in ADP fcis-12446 132 2 the the DET fcis-12446 132 3 detection detection NOUN fcis-12446 132 4 head head NOUN fcis-12446 132 5 section section NOUN fcis-12446 132 6 , , PUNCT fcis-12446 132 7 we we PRON fcis-12446 132 8 used use VERB fcis-12446 132 9 a a DET fcis-12446 132 10 cascading cascade VERB fcis-12446 132 11 approach approach NOUN fcis-12446 132 12 with with ADP fcis-12446 132 13 two two NUM fcis-12446 132 14 detection detection NOUN fcis-12446 132 15 heads head NOUN fcis-12446 132 16 for for ADP fcis-12446 132 17 detection detection NOUN fcis-12446 132 18 . . PUNCT fcis-12446 133 1 the the DET fcis-12446 133 2 input input NOUN fcis-12446 133 3 resolution resolution NOUN fcis-12446 133 4 of of ADP fcis-12446 133 5 images image NOUN fcis-12446 133 6 was be AUX fcis-12446 133 7 standardized standardize VERB fcis-12446 133 8 to to ADP fcis-12446 133 9 a a DET fcis-12446 133 10 certain certain ADJ fcis-12446 133 11 value value NOUN fcis-12446 133 12 . . PUNCT fcis-12446 134 1 table table NOUN fcis-12446 134 2 2 2 NUM fcis-12446 134 3 . . PUNCT fcis-12446 134 4 comparison comparison NOUN fcis-12446 134 5 of of ADP fcis-12446 134 6 experimental experimental ADJ fcis-12446 134 7 results result NOUN fcis-12446 134 8 model model PROPN fcis-12446 134 9 backbone backbone NOUN fcis-12446 134 10 ap ap PROPN fcis-12446 135 1 [ [ X fcis-12446 135 2 0.5:0.95 0.5:0.95 X fcis-12446 135 3 ] ] X fcis-12446 135 4 ap0.5 ap0.5 PRON fcis-12446 135 5 ap0.75 ap0.75 ADJ fcis-12446 135 6 input input NOUN fcis-12446 135 7 size size NOUN fcis-12446 135 8 faster fast ADV fcis-12446 135 9 - - PUNCT fcis-12446 135 10 rcnn rcnn NOUN fcis-12446 135 11 r101 r101 PROPN fcis-12446 135 12 0.190 0.190 NUM fcis-12446 135 13 0.305 0.305 NUM fcis-12446 135 14 0.190 0.190 NUM fcis-12446 135 15 1333×800 1333×800 NUM fcis-12446 135 16 r r NOUN fcis-12446 135 17 - - PUNCT fcis-12446 135 18 fcn fcn NOUN fcis-12446 135 19 r101 r101 PROPN fcis-12446 135 20 0.109 0.109 NUM fcis-12446 135 21 0.283 0.283 NUM fcis-12446 135 22 0.051 0.051 NUM fcis-12446 135 23 1333×800 1333×800 NUM fcis-12446 135 24 cascade cascade NOUN fcis-12446 135 25 - - PUNCT fcis-12446 135 26 rcnn rcnn NOUN fcis-12446 135 27 r101 r101 PROPN fcis-12446 135 28 0.183 0.183 NUM fcis-12446 135 29 0.306 0.306 NUM fcis-12446 135 30 0.178 0.178 NUM fcis-12446 135 31 1333×800 1333×800 NUM fcis-12446 135 32 ssd ssd NOUN fcis-12446 135 33 vgg16 vgg16 NOUN fcis-12446 135 34 0.136 0.136 NUM fcis-12446 135 35 0.277 0.277 NUM fcis-12446 135 36 0.086 0.086 NUM fcis-12446 135 37 512×512 512×512 NUM fcis-12446 135 38 retinanet retinanet NOUN fcis-12446 135 39 r101 r101 PROPN fcis-12446 135 40 0.174 0.174 NUM fcis-12446 135 41 0.298 0.298 NUM fcis-12446 135 42 0.161 0.161 NUM fcis-12446 135 43 1333×800 1333×800 NUM fcis-12446 135 44 yolov5 yolov5 NOUN fcis-12446 135 45 darknet53 darknet53 NOUN fcis-12446 136 1 0.237 0.237 NUM fcis-12446 136 2 0.526 0.526 NUM fcis-12446 136 3 0.207 0.207 NUM fcis-12446 136 4 640×640 640×640 NUM fcis-12446 136 5 fcos fco NOUN fcis-12446 136 6 hrnet hrnet NOUN fcis-12446 136 7 0.154 0.154 NUM fcis-12446 136 8 0.309 0.309 NUM fcis-12446 136 9 0.126 0.126 NUM fcis-12446 136 10 1333×800 1333×800 NUM fcis-12446 136 11 we we PRON fcis-12446 136 12 conducted conduct VERB fcis-12446 136 13 comparative comparative ADJ fcis-12446 136 14 experiments experiment NOUN fcis-12446 136 15 by by ADP fcis-12446 136 16 selecting select VERB fcis-12446 136 17 three three NUM fcis-12446 136 18 commonly commonly ADV fcis-12446 136 19 used use VERB fcis-12446 136 20 two two NUM fcis-12446 136 21 - - PUNCT fcis-12446 136 22 stage stage NOUN fcis-12446 136 23 object object NOUN fcis-12446 136 24 detection detection NOUN fcis-12446 136 25 algorithms algorithm NOUN fcis-12446 136 26 and and CCONJ fcis-12446 136 27 three three NUM fcis-12446 136 28 single single ADJ fcis-12446 136 29 - - PUNCT fcis-12446 136 30 stage stage NOUN fcis-12446 136 31 object object NOUN fcis-12446 136 32 detection detection NOUN fcis-12446 136 33 algorithms algorithm NOUN fcis-12446 136 34 , , PUNCT fcis-12446 136 35 along along ADP fcis-12446 136 36 with with ADP fcis-12446 136 37 one one NUM fcis-12446 136 38 anchor anchor NOUN fcis-12446 136 39 - - PUNCT fcis-12446 136 40 free free ADJ fcis-12446 136 41 method method NOUN fcis-12446 136 42 . . PUNCT fcis-12446 137 1 the the DET fcis-12446 137 2 selected select VERB fcis-12446 137 3 algorithms algorithm NOUN fcis-12446 137 4 include include VERB fcis-12446 137 5 faster fast ADJ fcis-12446 137 6 r r NOUN fcis-12446 137 7 - - PUNCT fcis-12446 137 8 cnn cnn PROPN fcis-12446 137 9 , , PUNCT fcis-12446 137 10 r r NOUN fcis-12446 137 11 - - PUNCT fcis-12446 137 12 fcn fcn NOUN fcis-12446 137 13 , , PUNCT fcis-12446 137 14 cascade cascade NOUN fcis-12446 137 15 r r NOUN fcis-12446 137 16 - - PUNCT fcis-12446 137 17 cnn cnn PROPN fcis-12446 137 18 , , PUNCT fcis-12446 137 19 ssd ssd PROPN fcis-12446 137 20 , , PUNCT fcis-12446 137 21 refinedet refinedet NOUN fcis-12446 137 22 , , PUNCT fcis-12446 137 23 yolov5 yolov5 NOUN fcis-12446 137 24 , , PUNCT fcis-12446 137 25 and and CCONJ fcis-12446 137 26 fcos fco NOUN fcis-12446 137 27 . . PUNCT fcis-12446 138 1 from from ADP fcis-12446 138 2 table table NOUN fcis-12446 138 3 2 2 NUM fcis-12446 138 4 , , PUNCT fcis-12446 138 5 we we PRON fcis-12446 138 6 can can AUX fcis-12446 138 7 observe observe VERB fcis-12446 138 8 that that SCONJ fcis-12446 138 9 our our PRON fcis-12446 138 10 os os ADJ fcis-12446 138 11 - - PUNCT fcis-12446 138 12 net net ADJ fcis-12446 138 13 network network NOUN fcis-12446 138 14 achieves achieve VERB fcis-12446 138 15 the the DET fcis-12446 138 16 best good ADJ fcis-12446 138 17 detection detection NOUN fcis-12446 138 18 results result NOUN fcis-12446 138 19 on on ADP fcis-12446 138 20 the the DET fcis-12446 138 21 ap ap PROPN fcis-12446 138 22 - - PUNCT fcis-12446 138 23 a a DET fcis-12446 138 24 dataset dataset NOUN fcis-12446 138 25 even even ADV fcis-12446 138 26 66 66 NUM fcis-12446 138 27 at at ADP fcis-12446 138 28 lower low ADJ fcis-12446 138 29 image image NOUN fcis-12446 138 30 resolutions resolution NOUN fcis-12446 138 31 , , PUNCT fcis-12446 138 32 further far ADV fcis-12446 138 33 demonstrating demonstrate VERB fcis-12446 138 34 the the DET fcis-12446 138 35 effectiveness effectiveness NOUN fcis-12446 138 36 of of ADP fcis-12446 138 37 the the DET fcis-12446 138 38 network network NOUN fcis-12446 138 39 architecture architecture NOUN fcis-12446 138 40 . . PUNCT fcis-12446 139 1 moreover moreover ADV fcis-12446 139 2 , , PUNCT fcis-12446 139 3 compared compare VERB fcis-12446 139 4 to to ADP fcis-12446 139 5 the the DET fcis-12446 139 6 baseline baseline PROPN fcis-12446 139 7 yolov5 yolov5 NOUN fcis-12446 139 8 network network NOUN fcis-12446 139 9 , , PUNCT fcis-12446 139 10 we we PRON fcis-12446 139 11 made make VERB fcis-12446 139 12 changes change NOUN fcis-12446 139 13 in in ADP fcis-12446 139 14 two two NUM fcis-12446 139 15 aspects aspect NOUN fcis-12446 139 16 : : PUNCT fcis-12446 139 17 the the DET fcis-12446 139 18 inclusion inclusion NOUN fcis-12446 139 19 of of ADP fcis-12446 139 20 an an DET fcis-12446 139 21 attention attention NOUN fcis-12446 139 22 mechanism mechanism NOUN fcis-12446 139 23 in in ADP fcis-12446 139 24 the the DET fcis-12446 139 25 backbone backbone NOUN fcis-12446 139 26 network network NOUN fcis-12446 139 27 and and CCONJ fcis-12446 139 28 the the DET fcis-12446 139 29 design design NOUN fcis-12446 139 30 of of ADP fcis-12446 139 31 an an DET fcis-12446 139 32 overall overall ADJ fcis-12446 139 33 - - PUNCT fcis-12446 139 34 to to ADP fcis-12446 139 35 - - PUNCT fcis-12446 139 36 specific specific ADJ fcis-12446 139 37 detection detection NOUN fcis-12446 139 38 head head NOUN fcis-12446 139 39 . . PUNCT fcis-12446 140 1 inserting insert VERB fcis-12446 140 2 four four NUM fcis-12446 140 3 ta ta NOUN fcis-12446 140 4 modules module NOUN fcis-12446 140 5 at at ADP fcis-12446 140 6 specified specified ADJ fcis-12446 140 7 locations location NOUN fcis-12446 140 8 in in ADP fcis-12446 140 9 the the DET fcis-12446 140 10 yolov5 yolov5 NOUN fcis-12446 140 11 backbone backbone NOUN fcis-12446 140 12 network network NOUN fcis-12446 140 13 improves improve VERB fcis-12446 140 14 the the DET fcis-12446 140 15 average average ADJ fcis-12446 140 16 accuracy accuracy NOUN fcis-12446 140 17 by by ADP fcis-12446 140 18 only only ADV fcis-12446 140 19 0.001 0.001 NUM fcis-12446 140 20 , , PUNCT fcis-12446 140 21 indicating indicate VERB fcis-12446 140 22 that that SCONJ fcis-12446 140 23 this this DET fcis-12446 140 24 improvement improvement NOUN fcis-12446 140 25 strategy strategy NOUN fcis-12446 140 26 has have VERB fcis-12446 140 27 a a DET fcis-12446 140 28 limited limited ADJ fcis-12446 140 29 impact impact NOUN fcis-12446 140 30 on on ADP fcis-12446 140 31 the the DET fcis-12446 140 32 final final ADJ fcis-12446 140 33 results result NOUN fcis-12446 140 34 . . PUNCT fcis-12446 141 1 in in ADP fcis-12446 141 2 other other ADJ fcis-12446 141 3 words word NOUN fcis-12446 141 4 , , PUNCT fcis-12446 141 5 the the DET fcis-12446 141 6 improvement improvement NOUN fcis-12446 141 7 in in ADP fcis-12446 141 8 average average ADJ fcis-12446 141 9 accuracy accuracy NOUN fcis-12446 141 10 of of ADP fcis-12446 141 11 the the DET fcis-12446 141 12 os os ADJ fcis-12446 141 13 - - PUNCT fcis-12446 141 14 net net ADJ fcis-12446 141 15 network network NOUN fcis-12446 141 16 primarily primarily ADV fcis-12446 141 17 stems stem VERB fcis-12446 141 18 from from ADP fcis-12446 141 19 the the DET fcis-12446 141 20 design design NOUN fcis-12446 141 21 of of ADP fcis-12446 141 22 the the DET fcis-12446 141 23 overall overall ADJ fcis-12446 141 24 - - PUNCT fcis-12446 141 25 to to ADP fcis-12446 141 26 - - PUNCT fcis-12446 141 27 specific specific ADJ fcis-12446 141 28 detection detection NOUN fcis-12446 141 29 head head NOUN fcis-12446 141 30 . . PUNCT fcis-12446 142 1 5 5 X fcis-12446 142 2 . . X fcis-12446 142 3 conclusion conclusion NOUN fcis-12446 142 4 we we PRON fcis-12446 142 5 proposed propose VERB fcis-12446 142 6 the the DET fcis-12446 142 7 os os ADJ fcis-12446 142 8 - - PUNCT fcis-12446 142 9 net net ADJ fcis-12446 142 10 network network NOUN fcis-12446 142 11 , , PUNCT fcis-12446 142 12 which which PRON fcis-12446 142 13 is be AUX fcis-12446 142 14 based base VERB fcis-12446 142 15 on on ADP fcis-12446 142 16 yolov5 yolov5 NOUN fcis-12446 142 17 and and CCONJ fcis-12446 142 18 incorporates incorporate VERB fcis-12446 142 19 our our PRON fcis-12446 142 20 concept concept NOUN fcis-12446 142 21 of of ADP fcis-12446 142 22 going go VERB fcis-12446 142 23 from from ADP fcis-12446 142 24 coarse coarse NOUN fcis-12446 142 25 to to ADP fcis-12446 142 26 fine fine NOUN fcis-12446 142 27 to to PART fcis-12446 142 28 achieve achieve VERB fcis-12446 142 29 aphid aphid NOUN fcis-12446 142 30 detection detection NOUN fcis-12446 142 31 . . PUNCT fcis-12446 143 1 specifically specifically ADV fcis-12446 143 2 , , PUNCT fcis-12446 143 3 we we PRON fcis-12446 143 4 added add VERB fcis-12446 143 5 an an DET fcis-12446 143 6 additional additional ADJ fcis-12446 143 7 detection detection NOUN fcis-12446 143 8 head head NOUN fcis-12446 143 9 , , PUNCT fcis-12446 143 10 where where SCONJ fcis-12446 143 11 the the DET fcis-12446 143 12 first first ADJ fcis-12446 143 13 head head NOUN fcis-12446 143 14 is be AUX fcis-12446 143 15 used use VERB fcis-12446 143 16 to to PART fcis-12446 143 17 detect detect VERB fcis-12446 143 18 densely densely ADV fcis-12446 143 19 annotated annotate VERB fcis-12446 143 20 aphid aphid NOUN fcis-12446 143 21 clusters cluster NOUN fcis-12446 143 22 , , PUNCT fcis-12446 143 23 and and CCONJ fcis-12446 143 24 the the DET fcis-12446 143 25 second second ADJ fcis-12446 143 26 component component NOUN fcis-12446 143 27 is be AUX fcis-12446 143 28 used use VERB fcis-12446 143 29 to to PART fcis-12446 143 30 capture capture VERB fcis-12446 143 31 more more ADV fcis-12446 143 32 small small ADJ fcis-12446 143 33 target target NOUN fcis-12446 143 34 aphids aphid NOUN fcis-12446 143 35 within within ADP fcis-12446 143 36 the the DET fcis-12446 143 37 target target NOUN fcis-12446 143 38 regions region NOUN fcis-12446 143 39 detected detect VERB fcis-12446 143 40 by by ADP fcis-12446 143 41 the the DET fcis-12446 143 42 first first ADJ fcis-12446 143 43 component component NOUN fcis-12446 143 44 by by ADP fcis-12446 143 45 matching match VERB fcis-12446 143 46 more more ADJ fcis-12446 143 47 anchor anchor NOUN fcis-12446 143 48 boxes box NOUN fcis-12446 143 49 . . PUNCT fcis-12446 144 1 the the DET fcis-12446 144 2 experiments experiment NOUN fcis-12446 144 3 demonstrate demonstrate VERB fcis-12446 144 4 the the DET fcis-12446 144 5 effectiveness effectiveness NOUN fcis-12446 144 6 of of ADP fcis-12446 144 7 our our PRON fcis-12446 144 8 approach approach NOUN fcis-12446 144 9 . . PUNCT fcis-12446 145 1 acknowledgments acknowledgment NOUN fcis-12446 145 2 this this DET fcis-12446 145 3 work work NOUN fcis-12446 145 4 was be AUX fcis-12446 145 5 supported support VERB fcis-12446 145 6 in in ADP fcis-12446 145 7 part part NOUN fcis-12446 145 8 by by ADP fcis-12446 145 9 a a DET fcis-12446 145 10 grant grant NOUN fcis-12446 145 11 from from ADP fcis-12446 145 12 zy2021002 zy2021002 PROPN fcis-12446 145 13 . . PUNCT fcis-12446 146 1 references reference NOUN fcis-12446 146 2 [ [ X fcis-12446 146 3 1 1 NUM fcis-12446 146 4 ] ] X fcis-12446 146 5 liu liu PROPN fcis-12446 146 6 j j PROPN fcis-12446 146 7 , , PUNCT fcis-12446 146 8 wang wang PROPN fcis-12446 146 9 x. x. PROPN fcis-12446 146 10 plant plant NOUN fcis-12446 146 11 diseases disease NOUN fcis-12446 146 12 and and CCONJ fcis-12446 146 13 pests pest NOUN fcis-12446 146 14 detection detection NOUN fcis-12446 146 15 based base VERB fcis-12446 146 16 on on ADP fcis-12446 146 17 deep deep ADJ fcis-12446 146 18 learning learning NOUN fcis-12446 146 19 : : PUNCT fcis-12446 146 20 a a DET fcis-12446 146 21 review[j review[j PROPN fcis-12446 146 22 ] ] PUNCT fcis-12446 146 23 . . PUNCT fcis-12446 147 1 plant plant NOUN fcis-12446 147 2 methods method NOUN fcis-12446 147 3 , , PUNCT fcis-12446 147 4 2021 2021 NUM fcis-12446 147 5 , , PUNCT fcis-12446 147 6 17 17 NUM fcis-12446 147 7 : : SYM fcis-12446 147 8 1 1 NUM fcis-12446 147 9 - - SYM fcis-12446 147 10 18 18 NUM fcis-12446 147 11 . . PUNCT fcis-12446 148 1 [ [ X fcis-12446 148 2 2 2 X fcis-12446 148 3 ] ] X fcis-12446 148 4 wolfert wolfert PROPN fcis-12446 148 5 s s PROPN fcis-12446 148 6 , , PUNCT fcis-12446 148 7 ge ge PROPN fcis-12446 148 8 l l PROPN fcis-12446 148 9 , , PUNCT fcis-12446 148 10 verdouw verdouw ADJ fcis-12446 148 11 c c X fcis-12446 148 12 , , PUNCT fcis-12446 148 13 et et PROPN fcis-12446 148 14 al al PROPN fcis-12446 148 15 . . PUNCT fcis-12446 149 1 big big ADJ fcis-12446 149 2 data datum NOUN fcis-12446 149 3 in in ADP fcis-12446 149 4 smart smart ADJ fcis-12446 149 5 farming farming NOUN fcis-12446 149 6 – – PUNCT fcis-12446 149 7 a a DET fcis-12446 149 8 review review NOUN fcis-12446 150 1 [ [ X fcis-12446 150 2 j j X fcis-12446 150 3 ] ] X fcis-12446 150 4 . . PUNCT fcis-12446 151 1 agricultural agricultural ADJ fcis-12446 151 2 systems system NOUN fcis-12446 151 3 , , PUNCT fcis-12446 151 4 2017 2017 NUM fcis-12446 151 5 , , PUNCT fcis-12446 151 6 153 153 NUM fcis-12446 151 7 : : PUNCT fcis-12446 151 8 69 69 NUM fcis-12446 151 9 - - SYM fcis-12446 151 10 80 80 NUM fcis-12446 151 11 . . PUNCT fcis-12446 152 1 [ [ X fcis-12446 152 2 3 3 X fcis-12446 152 3 ] ] X fcis-12446 152 4 girshick girshick ADJ fcis-12446 152 5 r r PROPN fcis-12446 152 6 , , PUNCT fcis-12446 152 7 donahue donahue PROPN fcis-12446 152 8 j j PROPN fcis-12446 152 9 , , PUNCT fcis-12446 152 10 darrell darrell PROPN fcis-12446 152 11 t t PROPN fcis-12446 152 12 , , PUNCT fcis-12446 152 13 et et PROPN fcis-12446 152 14 al al PROPN fcis-12446 152 15 . . PROPN fcis-12446 152 16 rich rich ADJ fcis-12446 152 17 feature feature NOUN fcis-12446 152 18 hierarchies hierarchy NOUN fcis-12446 152 19 for for ADP fcis-12446 152 20 accurate accurate ADJ fcis-12446 152 21 object object NOUN fcis-12446 152 22 detection detection NOUN fcis-12446 152 23 and and CCONJ fcis-12446 152 24 semantic semantic ADJ fcis-12446 152 25 segmentation segmentation NOUN fcis-12446 152 26 [ [ X fcis-12446 152 27 c]// c]// ADJ fcis-12446 152 28 proceedings proceeding NOUN fcis-12446 152 29 of of ADP fcis-12446 152 30 the the DET fcis-12446 152 31 ieee ieee NOUN fcis-12446 152 32 conference conference NOUN fcis-12446 152 33 on on ADP fcis-12446 152 34 computer computer NOUN fcis-12446 152 35 vision vision NOUN fcis-12446 152 36 and and CCONJ fcis-12446 152 37 pattern pattern NOUN fcis-12446 152 38 recognition recognition NOUN fcis-12446 152 39 . . PUNCT fcis-12446 153 1 2014 2014 NUM fcis-12446 153 2 : : PUNCT fcis-12446 154 1 580 580 NUM fcis-12446 154 2 - - SYM fcis-12446 154 3 587 587 NUM fcis-12446 154 4 . . PUNCT fcis-12446 155 1 [ [ X fcis-12446 155 2 4 4 X fcis-12446 155 3 ] ] X fcis-12446 155 4 yaakob yaakob PROPN fcis-12446 155 5 s s PART fcis-12446 155 6 n n CCONJ fcis-12446 155 7 , , PUNCT fcis-12446 155 8 jain jain PROPN fcis-12446 155 9 l. l. PROPN fcis-12446 155 10 an an DET fcis-12446 155 11 insect insect NOUN fcis-12446 155 12 classification classification NOUN fcis-12446 155 13 analysis analysis NOUN fcis-12446 155 14 based base VERB fcis-12446 155 15 on on ADP fcis-12446 155 16 shape shape NOUN fcis-12446 155 17 features feature NOUN fcis-12446 155 18 using use VERB fcis-12446 155 19 quality quality NOUN fcis-12446 155 20 threshold threshold NOUN fcis-12446 155 21 artmap artmap NOUN fcis-12446 155 22 and and CCONJ fcis-12446 155 23 moment moment VERB fcis-12446 155 24 invariant[j invariant[j ADV fcis-12446 155 25 ] ] PUNCT fcis-12446 155 26 . . PUNCT fcis-12446 156 1 applied apply VERB fcis-12446 156 2 intelligence intelligence NOUN fcis-12446 156 3 , , PUNCT fcis-12446 156 4 2012 2012 NUM fcis-12446 156 5 , , PUNCT fcis-12446 156 6 37 37 NUM fcis-12446 156 7 : : SYM fcis-12446 156 8 12 12 NUM fcis-12446 156 9 - - SYM fcis-12446 156 10 30 30 NUM fcis-12446 156 11 . . PUNCT fcis-12446 157 1 [ [ X fcis-12446 157 2 5 5 X fcis-12446 157 3 ] ] PUNCT fcis-12446 157 4 wang wang PROPN fcis-12446 157 5 j j PROPN fcis-12446 157 6 , , PUNCT fcis-12446 157 7 lin lin PROPN fcis-12446 157 8 c c PROPN fcis-12446 157 9 , , PUNCT fcis-12446 157 10 ji ji PROPN fcis-12446 157 11 l l PROPN fcis-12446 157 12 , , PUNCT fcis-12446 157 13 et et PROPN fcis-12446 157 14 al al PROPN fcis-12446 157 15 . . PUNCT fcis-12446 158 1 a a DET fcis-12446 158 2 new new ADJ fcis-12446 158 3 automatic automatic ADJ fcis-12446 158 4 identification identification NOUN fcis-12446 158 5 system system NOUN fcis-12446 158 6 of of ADP fcis-12446 158 7 insect insect NOUN fcis-12446 158 8 images image NOUN fcis-12446 158 9 at at ADP fcis-12446 158 10 the the DET fcis-12446 158 11 order order NOUN fcis-12446 158 12 level[j level[j X fcis-12446 158 13 ] ] PUNCT fcis-12446 158 14 . . PUNCT fcis-12446 159 1 knowledge knowledge NOUN fcis-12446 159 2 - - PUNCT fcis-12446 159 3 based base VERB fcis-12446 159 4 systems system NOUN fcis-12446 159 5 , , PUNCT fcis-12446 159 6 2012 2012 NUM fcis-12446 159 7 , , PUNCT fcis-12446 159 8 33 33 NUM fcis-12446 159 9 : : SYM fcis-12446 159 10 102 102 NUM fcis-12446 159 11 - - SYM fcis-12446 159 12 110 110 NUM fcis-12446 159 13 . . PUNCT fcis-12446 160 1 [ [ X fcis-12446 160 2 6 6 NUM fcis-12446 160 3 ] ] PUNCT fcis-12446 160 4 qing qe VERB fcis-12446 160 5 y y PROPN fcis-12446 160 6 a a DET fcis-12446 160 7 o o NOUN fcis-12446 160 8 , , PUNCT fcis-12446 160 9 jun jun PROPN fcis-12446 160 10 l l PROPN fcis-12446 160 11 v v PROPN fcis-12446 160 12 , , PUNCT fcis-12446 160 13 liu liu PROPN fcis-12446 160 14 q q PROPN fcis-12446 160 15 , , PUNCT fcis-12446 160 16 et et PROPN fcis-12446 160 17 al al PROPN fcis-12446 160 18 . . PUNCT fcis-12446 161 1 an an DET fcis-12446 161 2 insect insect NOUN fcis-12446 161 3 imaging imaging NOUN fcis-12446 161 4 system system NOUN fcis-12446 161 5 to to PART fcis-12446 161 6 automate automate VERB fcis-12446 161 7 rice rice NOUN fcis-12446 161 8 light light NOUN fcis-12446 161 9 - - PUNCT fcis-12446 161 10 trap trap NOUN fcis-12446 161 11 pest pest NOUN fcis-12446 161 12 identification[j identification[j PROPN fcis-12446 161 13 ] ] PUNCT fcis-12446 161 14 . . PUNCT fcis-12446 162 1 journal journal PROPN fcis-12446 162 2 of of ADP fcis-12446 162 3 integrative integrative ADJ fcis-12446 162 4 agriculture agriculture NOUN fcis-12446 162 5 , , PUNCT fcis-12446 162 6 2012 2012 NUM fcis-12446 162 7 , , PUNCT fcis-12446 162 8 11(6 11(6 NUM fcis-12446 162 9 ): ): PUNCT fcis-12446 162 10 978 978 NUM fcis-12446 162 11 - - SYM fcis-12446 162 12 985 985 NUM fcis-12446 162 13 . . PUNCT fcis-12446 163 1 [ [ X fcis-12446 163 2 7 7 NUM fcis-12446 163 3 ] ] X fcis-12446 163 4 tian tian ADJ fcis-12446 163 5 h h NOUN fcis-12446 163 6 , , PUNCT fcis-12446 163 7 wang wang PROPN fcis-12446 163 8 t t PROPN fcis-12446 163 9 , , PUNCT fcis-12446 163 10 liu liu PROPN fcis-12446 163 11 y y PROPN fcis-12446 163 12 , , PUNCT fcis-12446 163 13 et et PROPN fcis-12446 164 1 al al PROPN fcis-12446 164 2 . . PUNCT fcis-12446 164 3 computer computer NOUN fcis-12446 164 4 vision vision NOUN fcis-12446 164 5 technology technology NOUN fcis-12446 164 6 in in ADP fcis-12446 164 7 agricultural agricultural ADJ fcis-12446 164 8 automation automation NOUN fcis-12446 164 9 -- -- PUNCT fcis-12446 164 10 a a DET fcis-12446 164 11 review[j review[j NOUN fcis-12446 164 12 ] ] PUNCT fcis-12446 164 13 . . PUNCT fcis-12446 165 1 information information NOUN fcis-12446 165 2 processing processing NOUN fcis-12446 165 3 in in ADP fcis-12446 165 4 agriculture agriculture NOUN fcis-12446 165 5 , , PUNCT fcis-12446 165 6 2020 2020 NUM fcis-12446 165 7 , , PUNCT fcis-12446 165 8 7(1 7(1 NUM fcis-12446 165 9 ): ): PUNCT fcis-12446 165 10 1 1 NUM fcis-12446 165 11 - - SYM fcis-12446 165 12 19 19 NUM fcis-12446 165 13 . . PUNCT fcis-12446 166 1 [ [ X fcis-12446 166 2 8 8 NUM fcis-12446 166 3 ] ] X fcis-12446 166 4 al al PROPN fcis-12446 166 5 ohali ohali PROPN fcis-12446 166 6 y. y. PROPN fcis-12446 166 7 computer computer PROPN fcis-12446 166 8 vision vision PROPN fcis-12446 166 9 based base VERB fcis-12446 166 10 date date NOUN fcis-12446 166 11 fruit fruit NOUN fcis-12446 166 12 grading grade VERB fcis-12446 166 13 system system NOUN fcis-12446 166 14 : : PUNCT fcis-12446 166 15 design design NOUN fcis-12446 166 16 and and CCONJ fcis-12446 166 17 implementation[j implementation[j NOUN fcis-12446 166 18 ] ] PUNCT fcis-12446 166 19 . . PUNCT fcis-12446 167 1 journal journal PROPN fcis-12446 167 2 of of ADP fcis-12446 167 3 king king PROPN fcis-12446 167 4 saud saud PROPN fcis-12446 167 5 university university PROPN fcis-12446 167 6 - - PUNCT fcis-12446 167 7 computer computer NOUN fcis-12446 167 8 and and CCONJ fcis-12446 167 9 information information NOUN fcis-12446 167 10 sciences science NOUN fcis-12446 167 11 , , PUNCT fcis-12446 167 12 2011 2011 NUM fcis-12446 167 13 , , PUNCT fcis-12446 167 14 23(1 23(1 NUM fcis-12446 167 15 ): ): PUNCT fcis-12446 167 16 29 29 NUM fcis-12446 167 17 - - SYM fcis-12446 167 18 36 36 NUM fcis-12446 167 19 . . PUNCT fcis-12446 168 1 [ [ X fcis-12446 168 2 9 9 NUM fcis-12446 168 3 ] ] SYM fcis-12446 168 4 weeks week NOUN fcis-12446 169 1 p p X fcis-12446 169 2 j j PROPN fcis-12446 169 3 d d PROPN fcis-12446 169 4 , , PUNCT fcis-12446 169 5 o'neill o'neill PROPN fcis-12446 169 6 m m PROPN fcis-12446 169 7 a a PROPN fcis-12446 169 8 , , PUNCT fcis-12446 169 9 gaston gaston PROPN fcis-12446 169 10 k k PROPN fcis-12446 169 11 j j PROPN fcis-12446 169 12 , , PUNCT fcis-12446 169 13 et et PROPN fcis-12446 169 14 al al PROPN fcis-12446 169 15 . . PUNCT fcis-12446 169 16 species species PROPN fcis-12446 169 17 – – PUNCT fcis-12446 169 18 identification identification NOUN fcis-12446 169 19 of of ADP fcis-12446 169 20 wasps wasp NOUN fcis-12446 169 21 using use VERB fcis-12446 169 22 principal principal ADJ fcis-12446 169 23 component component NOUN fcis-12446 169 24 associative associative NOUN fcis-12446 169 25 memories memory NOUN fcis-12446 170 1 [ [ X fcis-12446 170 2 j j X fcis-12446 170 3 ] ] X fcis-12446 170 4 . . PUNCT fcis-12446 171 1 image image NOUN fcis-12446 171 2 and and CCONJ fcis-12446 171 3 vision vision NOUN fcis-12446 171 4 computing computing NOUN fcis-12446 171 5 , , PUNCT fcis-12446 171 6 1999 1999 NUM fcis-12446 171 7 , , PUNCT fcis-12446 171 8 17(12 17(12 NUM fcis-12446 171 9 ): ): PUNCT fcis-12446 171 10 861 861 NUM fcis-12446 171 11 - - SYM fcis-12446 171 12 866 866 NUM fcis-12446 171 13 . . PUNCT fcis-12446 172 1 [ [ X fcis-12446 172 2 10 10 NUM fcis-12446 172 3 ] ] X fcis-12446 172 4 cho cho PROPN fcis-12446 172 5 j j PROPN fcis-12446 172 6 , , PUNCT fcis-12446 172 7 choi choi PROPN fcis-12446 172 8 j j PROPN fcis-12446 172 9 , , PUNCT fcis-12446 172 10 qiao qiao PROPN fcis-12446 172 11 m m PROPN fcis-12446 172 12 , , PUNCT fcis-12446 172 13 et et PROPN fcis-12446 172 14 al al PROPN fcis-12446 172 15 . . PROPN fcis-12446 172 16 automatic automatic ADJ fcis-12446 172 17 identification identification NOUN fcis-12446 172 18 of of ADP fcis-12446 172 19 whiteflies whitefly NOUN fcis-12446 172 20 , , PUNCT fcis-12446 172 21 aphids aphid NOUN fcis-12446 172 22 and and CCONJ fcis-12446 172 23 thrips thrip NOUN fcis-12446 172 24 in in ADP fcis-12446 172 25 greenhouse greenhouse NOUN fcis-12446 172 26 based base VERB fcis-12446 172 27 on on ADP fcis-12446 172 28 image image NOUN fcis-12446 172 29 analysis analysis NOUN fcis-12446 173 1 [ [ X fcis-12446 173 2 j j X fcis-12446 173 3 ] ] X fcis-12446 173 4 . . PUNCT fcis-12446 174 1 red red ADJ fcis-12446 174 2 , , PUNCT fcis-12446 174 3 2007 2007 NUM fcis-12446 174 4 , , PUNCT fcis-12446 174 5 346(246 346(246 NUM fcis-12446 174 6 ): ): PUNCT fcis-12446 174 7 244 244 NUM fcis-12446 174 8 . . PUNCT fcis-12446 175 1 [ [ X fcis-12446 175 2 11 11 NUM fcis-12446 175 3 ] ] X fcis-12446 175 4 solis solis PROPN fcis-12446 175 5 - - PUNCT fcis-12446 175 6 sánchez sánchez PROPN fcis-12446 175 7 l l PROPN fcis-12446 175 8 o o NOUN fcis-12446 175 9 , , PUNCT fcis-12446 175 10 castañeda castañeda PROPN fcis-12446 175 11 - - PUNCT fcis-12446 175 12 miranda miranda NOUN fcis-12446 175 13 r r NOUN fcis-12446 175 14 , , PUNCT fcis-12446 175 15 garcía garcía NOUN fcis-12446 175 16 - - PUNCT fcis-12446 175 17 escalante escalante NOUN fcis-12446 175 18 j j PROPN fcis-12446 175 19 j j PROPN fcis-12446 175 20 , , PUNCT fcis-12446 175 21 et et PROPN fcis-12446 175 22 al al PROPN fcis-12446 175 23 . . PROPN fcis-12446 175 24 scale scale PROPN fcis-12446 175 25 invariant invariant ADJ fcis-12446 175 26 feature feature NOUN fcis-12446 175 27 approach approach NOUN fcis-12446 175 28 for for ADP fcis-12446 175 29 insect insect NOUN fcis-12446 175 30 monitoring monitor VERB fcis-12446 175 31 [ [ X fcis-12446 175 32 j j X fcis-12446 175 33 ] ] X fcis-12446 175 34 . . PUNCT fcis-12446 176 1 computers computer NOUN fcis-12446 176 2 and and CCONJ fcis-12446 176 3 electronics electronic NOUN fcis-12446 176 4 in in ADP fcis-12446 176 5 agriculture agriculture NOUN fcis-12446 176 6 , , PUNCT fcis-12446 176 7 2011 2011 NUM fcis-12446 176 8 , , PUNCT fcis-12446 176 9 75(1 75(1 NUM fcis-12446 176 10 ): ): PUNCT fcis-12446 176 11 9299 9299 NUM fcis-12446 176 12 . . PUNCT