id sid tid token lemma pos fcis-7187 1 1 frontiers frontier NOUN fcis-7187 1 2 in in ADP fcis-7187 1 3 computing computing NOUN fcis-7187 1 4 and and CCONJ fcis-7187 1 5 intelligent intelligent ADJ fcis-7187 1 6 systems system NOUN fcis-7187 1 7 issn issn VERB fcis-7187 1 8 : : PUNCT fcis-7187 1 9 2832 2832 NUM fcis-7187 1 10 - - SYM fcis-7187 1 11 6024 6024 NUM fcis-7187 1 12 | | NOUN fcis-7187 1 13 vol vol NOUN fcis-7187 1 14 . . PROPN fcis-7187 2 1 3 3 NUM fcis-7187 2 2 , , PUNCT fcis-7187 2 3 no no INTJ fcis-7187 2 4 . . NOUN fcis-7187 2 5 2 2 NUM fcis-7187 2 6 , , PUNCT fcis-7187 2 7 2023 2023 NUM fcis-7187 3 1 48 48 NUM fcis-7187 3 2 tea tea NOUN fcis-7187 3 3 leaf leaf NOUN fcis-7187 3 4 disease disease NOUN fcis-7187 3 5 classification classification NOUN fcis-7187 3 6 using use VERB fcis-7187 3 7 domain domain NOUN fcis-7187 3 8 adaptation adaptation NOUN fcis-7187 3 9 method method NOUN fcis-7187 3 10 wei wei PROPN fcis-7187 3 11 wu wu PROPN fcis-7187 3 12 taishan taishan PROPN fcis-7187 3 13 university university PROPN fcis-7187 3 14 , , PUNCT fcis-7187 3 15 tai'an tai'an PROPN fcis-7187 3 16 271000 271000 NUM fcis-7187 3 17 , , PUNCT fcis-7187 3 18 china china PROPN fcis-7187 3 19 abstract abstract NOUN fcis-7187 3 20 : : PUNCT fcis-7187 3 21 tea tea NOUN fcis-7187 3 22 trees tree NOUN fcis-7187 3 23 are be AUX fcis-7187 3 24 extremely extremely ADV fcis-7187 3 25 vulnerable vulnerable ADJ fcis-7187 3 26 to to ADP fcis-7187 3 27 diseases disease NOUN fcis-7187 3 28 and and CCONJ fcis-7187 3 29 insect insect VERB fcis-7187 3 30 pests pest NOUN fcis-7187 3 31 in in ADP fcis-7187 3 32 the the DET fcis-7187 3 33 growth growth NOUN fcis-7187 3 34 process process NOUN fcis-7187 3 35 , , PUNCT fcis-7187 3 36 which which PRON fcis-7187 3 37 seriously seriously ADV fcis-7187 3 38 affects affect VERB fcis-7187 3 39 the the DET fcis-7187 3 40 yield yield NOUN fcis-7187 3 41 and and CCONJ fcis-7187 3 42 quality quality NOUN fcis-7187 3 43 of of ADP fcis-7187 3 44 tea tea NOUN fcis-7187 3 45 leaves leave NOUN fcis-7187 3 46 . . PUNCT fcis-7187 4 1 this this PRON fcis-7187 4 2 requires require VERB fcis-7187 4 3 the the DET fcis-7187 4 4 identification identification NOUN fcis-7187 4 5 and and CCONJ fcis-7187 4 6 treatment treatment NOUN fcis-7187 4 7 of of ADP fcis-7187 4 8 infected infected ADJ fcis-7187 4 9 tea tea NOUN fcis-7187 4 10 leaves leave NOUN fcis-7187 4 11 in in ADP fcis-7187 4 12 time time NOUN fcis-7187 4 13 . . PUNCT fcis-7187 5 1 with with ADP fcis-7187 5 2 the the DET fcis-7187 5 3 rapid rapid ADJ fcis-7187 5 4 development development NOUN fcis-7187 5 5 of of ADP fcis-7187 5 6 artificial artificial ADJ fcis-7187 5 7 intelligence intelligence NOUN fcis-7187 5 8 and and CCONJ fcis-7187 5 9 computer computer NOUN fcis-7187 5 10 vision vision NOUN fcis-7187 5 11 , , PUNCT fcis-7187 5 12 there there PRON fcis-7187 5 13 is be VERB fcis-7187 5 14 a a DET fcis-7187 5 15 new new ADJ fcis-7187 5 16 trend trend NOUN fcis-7187 5 17 to to PART fcis-7187 5 18 identify identify VERB fcis-7187 5 19 the the DET fcis-7187 5 20 images image NOUN fcis-7187 5 21 by by ADP fcis-7187 5 22 computer computer NOUN fcis-7187 5 23 . . PUNCT fcis-7187 6 1 however however ADV fcis-7187 6 2 , , PUNCT fcis-7187 6 3 there there PRON fcis-7187 6 4 often often ADV fcis-7187 6 5 exists exist VERB fcis-7187 6 6 environmental environmental ADJ fcis-7187 6 7 changes change NOUN fcis-7187 6 8 such such ADJ fcis-7187 6 9 as as ADP fcis-7187 6 10 illumination illumination NOUN fcis-7187 6 11 intensity intensity NOUN fcis-7187 6 12 , , PUNCT fcis-7187 6 13 sample sample NOUN fcis-7187 6 14 angles angle NOUN fcis-7187 6 15 and and CCONJ fcis-7187 6 16 background background NOUN fcis-7187 6 17 in in ADP fcis-7187 6 18 the the DET fcis-7187 6 19 leaf leaf NOUN fcis-7187 6 20 images image NOUN fcis-7187 6 21 collected collect VERB fcis-7187 6 22 in in ADP fcis-7187 6 23 natural natural ADJ fcis-7187 6 24 scenes scene NOUN fcis-7187 6 25 , , PUNCT fcis-7187 6 26 which which PRON fcis-7187 6 27 results result VERB fcis-7187 6 28 in in ADP fcis-7187 6 29 the the DET fcis-7187 6 30 difference difference NOUN fcis-7187 6 31 of of ADP fcis-7187 6 32 data datum NOUN fcis-7187 6 33 distribution distribution NOUN fcis-7187 6 34 . . PUNCT fcis-7187 7 1 this this PRON fcis-7187 7 2 makes make VERB fcis-7187 7 3 the the DET fcis-7187 7 4 traditional traditional ADJ fcis-7187 7 5 deep deep ADJ fcis-7187 7 6 learning learning NOUN fcis-7187 7 7 method method NOUN fcis-7187 7 8 unable unable ADJ fcis-7187 7 9 to to PART fcis-7187 7 10 solve solve VERB fcis-7187 7 11 the the DET fcis-7187 7 12 problem problem NOUN fcis-7187 7 13 of of ADP fcis-7187 7 14 cross cross ADJ fcis-7187 7 15 - - ADJ fcis-7187 7 16 domain domain ADJ fcis-7187 7 17 classification classification NOUN fcis-7187 7 18 very very ADV fcis-7187 7 19 well well ADV fcis-7187 7 20 , , PUNCT fcis-7187 7 21 thus thus ADV fcis-7187 7 22 seriously seriously ADV fcis-7187 7 23 affecting affect VERB fcis-7187 7 24 the the DET fcis-7187 7 25 accuracy accuracy NOUN fcis-7187 7 26 of of ADP fcis-7187 7 27 classification classification NOUN fcis-7187 7 28 . . PUNCT fcis-7187 8 1 this this DET fcis-7187 8 2 study study NOUN fcis-7187 8 3 mainly mainly ADV fcis-7187 8 4 focuses focus VERB fcis-7187 8 5 on on ADP fcis-7187 8 6 the the DET fcis-7187 8 7 cross cross ADJ fcis-7187 8 8 - - ADJ fcis-7187 8 9 domain domain ADJ fcis-7187 8 10 task task NOUN fcis-7187 8 11 with with ADP fcis-7187 8 12 unaligned unaligned ADJ fcis-7187 8 13 data datum NOUN fcis-7187 8 14 distribution distribution NOUN fcis-7187 8 15 . . PUNCT fcis-7187 9 1 a a DET fcis-7187 9 2 domain domain NOUN fcis-7187 9 3 - - PUNCT fcis-7187 9 4 adaptive adaptive NOUN fcis-7187 9 5 based base VERB fcis-7187 9 6 method method NOUN fcis-7187 9 7 was be AUX fcis-7187 9 8 proposed propose VERB fcis-7187 9 9 to to PART fcis-7187 9 10 realize realize VERB fcis-7187 9 11 the the DET fcis-7187 9 12 cross cross ADJ fcis-7187 9 13 - - ADJ fcis-7187 9 14 domain domain ADJ fcis-7187 9 15 classification classification NOUN fcis-7187 9 16 of of ADP fcis-7187 9 17 tea tea NOUN fcis-7187 9 18 leaf leaf NOUN fcis-7187 9 19 diseases disease NOUN fcis-7187 9 20 . . PUNCT fcis-7187 10 1 the the DET fcis-7187 10 2 experimental experimental ADJ fcis-7187 10 3 results result NOUN fcis-7187 10 4 verify verify VERB fcis-7187 10 5 the the DET fcis-7187 10 6 effectiveness effectiveness NOUN fcis-7187 10 7 of of ADP fcis-7187 10 8 the the DET fcis-7187 10 9 presented present VERB fcis-7187 10 10 method method NOUN fcis-7187 10 11 and and CCONJ fcis-7187 10 12 provide provide VERB fcis-7187 10 13 new new ADJ fcis-7187 10 14 thoughts thought NOUN fcis-7187 10 15 for for ADP fcis-7187 10 16 the the DET fcis-7187 10 17 cross cross ADJ fcis-7187 10 18 - - ADJ fcis-7187 10 19 domain domain ADJ fcis-7187 10 20 classification classification NOUN fcis-7187 10 21 problem problem NOUN fcis-7187 10 22 in in ADP fcis-7187 10 23 agriculture agriculture NOUN fcis-7187 10 24 . . PUNCT fcis-7187 11 1 keywords keyword NOUN fcis-7187 11 2 : : PUNCT fcis-7187 11 3 deep deep ADJ fcis-7187 11 4 learning learning NOUN fcis-7187 11 5 ; ; PUNCT fcis-7187 11 6 domain domain NOUN fcis-7187 11 7 adaptation adaptation NOUN fcis-7187 11 8 ; ; PUNCT fcis-7187 11 9 tea tea NOUN fcis-7187 11 10 leaf leaf NOUN fcis-7187 11 11 disease disease NOUN fcis-7187 11 12 classification classification NOUN fcis-7187 11 13 . . PUNCT fcis-7187 12 1 1 1 X fcis-7187 12 2 . . X fcis-7187 12 3 introduction introduction NOUN fcis-7187 12 4 tea tea NOUN fcis-7187 12 5 tree tree NOUN fcis-7187 12 6 is be AUX fcis-7187 12 7 an an DET fcis-7187 12 8 important important ADJ fcis-7187 12 9 economic economic ADJ fcis-7187 12 10 crop crop NOUN fcis-7187 12 11 in in ADP fcis-7187 12 12 the the DET fcis-7187 12 13 world world NOUN fcis-7187 12 14 and and CCONJ fcis-7187 12 15 an an DET fcis-7187 12 16 important important ADJ fcis-7187 12 17 industry industry NOUN fcis-7187 12 18 to to PART fcis-7187 12 19 drive drive VERB fcis-7187 12 20 tea tea NOUN fcis-7187 12 21 farmers farmer NOUN fcis-7187 12 22 out out ADP fcis-7187 12 23 of of ADP fcis-7187 12 24 poverty poverty NOUN fcis-7187 12 25 . . PUNCT fcis-7187 13 1 however however ADV fcis-7187 13 2 , , PUNCT fcis-7187 13 3 tea tea NOUN fcis-7187 13 4 leaves leave NOUN fcis-7187 13 5 are be AUX fcis-7187 13 6 often often ADV fcis-7187 13 7 attacked attack VERB fcis-7187 13 8 by by ADP fcis-7187 13 9 diseases disease NOUN fcis-7187 13 10 and and CCONJ fcis-7187 13 11 insect insect NOUN fcis-7187 13 12 pests pest NOUN fcis-7187 13 13 . . PUNCT fcis-7187 14 1 it it PRON fcis-7187 14 2 is be AUX fcis-7187 14 3 of of ADP fcis-7187 14 4 great great ADJ fcis-7187 14 5 significance significance NOUN fcis-7187 14 6 to to PART fcis-7187 14 7 ensure ensure VERB fcis-7187 14 8 the the DET fcis-7187 14 9 yield yield NOUN fcis-7187 14 10 and and CCONJ fcis-7187 14 11 quality quality NOUN fcis-7187 14 12 of of ADP fcis-7187 14 13 tea tea NOUN fcis-7187 14 14 by by ADP fcis-7187 14 15 identifying identify VERB fcis-7187 14 16 the the DET fcis-7187 14 17 types type NOUN fcis-7187 14 18 of of ADP fcis-7187 14 19 infected infected ADJ fcis-7187 14 20 tea tea NOUN fcis-7187 14 21 leaves leave NOUN fcis-7187 14 22 accurately accurately ADV fcis-7187 14 23 and and CCONJ fcis-7187 14 24 taking take VERB fcis-7187 14 25 protective protective ADJ fcis-7187 14 26 measures measure NOUN fcis-7187 14 27 timely timely ADV fcis-7187 14 28 . . PUNCT fcis-7187 15 1 the the DET fcis-7187 15 2 appearance appearance NOUN fcis-7187 15 3 and and CCONJ fcis-7187 15 4 growth growth NOUN fcis-7187 15 5 state state NOUN fcis-7187 15 6 of of ADP fcis-7187 15 7 tea tea NOUN fcis-7187 15 8 leaves leave NOUN fcis-7187 15 9 can can AUX fcis-7187 15 10 provide provide VERB fcis-7187 15 11 a a DET fcis-7187 15 12 prediction prediction NOUN fcis-7187 15 13 basis basis NOUN fcis-7187 15 14 for for ADP fcis-7187 15 15 the the DET fcis-7187 15 16 growth growth NOUN fcis-7187 15 17 trend trend NOUN fcis-7187 15 18 of of ADP fcis-7187 15 19 tea tea NOUN fcis-7187 15 20 trees tree NOUN fcis-7187 15 21 . . PUNCT fcis-7187 16 1 with with ADP fcis-7187 16 2 the the DET fcis-7187 16 3 rapid rapid ADJ fcis-7187 16 4 development development NOUN fcis-7187 16 5 of of ADP fcis-7187 16 6 computer computer NOUN fcis-7187 16 7 technology technology NOUN fcis-7187 16 8 , , PUNCT fcis-7187 16 9 the the DET fcis-7187 16 10 identification identification NOUN fcis-7187 16 11 of of ADP fcis-7187 16 12 tea tea NOUN fcis-7187 16 13 leaves leave NOUN fcis-7187 16 14 images image NOUN fcis-7187 16 15 has have AUX fcis-7187 16 16 been be AUX fcis-7187 16 17 widely widely ADV fcis-7187 16 18 studied study VERB fcis-7187 16 19 with with ADP fcis-7187 16 20 the the DET fcis-7187 16 21 help help NOUN fcis-7187 16 22 of of ADP fcis-7187 16 23 computer computer NOUN fcis-7187 16 24 vision vision NOUN fcis-7187 16 25 . . PUNCT fcis-7187 17 1 since since SCONJ fcis-7187 17 2 2006 2006 NUM fcis-7187 17 3 , , PUNCT fcis-7187 17 4 deep deep ADJ fcis-7187 17 5 learning learning NOUN fcis-7187 17 6 technology technology NOUN fcis-7187 17 7 has have AUX fcis-7187 17 8 developed develop VERB fcis-7187 17 9 rapidly rapidly ADV fcis-7187 17 10 and and CCONJ fcis-7187 17 11 has have AUX fcis-7187 17 12 made make VERB fcis-7187 17 13 remarkable remarkable ADJ fcis-7187 17 14 achievements achievement NOUN fcis-7187 17 15 in in ADP fcis-7187 17 16 the the DET fcis-7187 17 17 field field NOUN fcis-7187 17 18 of of ADP fcis-7187 17 19 image image NOUN fcis-7187 17 20 recognition recognition NOUN fcis-7187 17 21 . . PUNCT fcis-7187 18 1 as as ADP fcis-7187 18 2 the the DET fcis-7187 18 3 core core NOUN fcis-7187 18 4 technology technology NOUN fcis-7187 18 5 of of ADP fcis-7187 18 6 deep deep ADJ fcis-7187 18 7 learning learning NOUN fcis-7187 18 8 , , PUNCT fcis-7187 18 9 convolutional convolutional ADJ fcis-7187 18 10 neural neural ADJ fcis-7187 18 11 network network NOUN fcis-7187 18 12 ( ( PUNCT fcis-7187 18 13 cnn cnn PROPN fcis-7187 18 14 ) ) PUNCT fcis-7187 18 15 is be AUX fcis-7187 18 16 widely widely ADV fcis-7187 18 17 used use VERB fcis-7187 18 18 in in ADP fcis-7187 18 19 the the DET fcis-7187 18 20 field field NOUN fcis-7187 18 21 of of ADP fcis-7187 18 22 automatic automatic ADJ fcis-7187 18 23 recognition recognition NOUN fcis-7187 18 24 of of ADP fcis-7187 18 25 tea tea NOUN fcis-7187 18 26 leaf leaf NOUN fcis-7187 18 27 images image NOUN fcis-7187 18 28 . . PUNCT fcis-7187 19 1 chen chen PROPN fcis-7187 19 2 et et PROPN fcis-7187 19 3 al al PROPN fcis-7187 19 4 . . PUNCT fcis-7187 20 1 [ [ X fcis-7187 20 2 1 1 X fcis-7187 20 3 ] ] PUNCT fcis-7187 20 4 developed develop VERB fcis-7187 20 5 a a DET fcis-7187 20 6 cnn cnn PROPN fcis-7187 20 7 model model NOUN fcis-7187 20 8 called call VERB fcis-7187 20 9 leafnet leafnet PROPN fcis-7187 20 10 to to PART fcis-7187 20 11 classify classify VERB fcis-7187 20 12 tea tea NOUN fcis-7187 20 13 leaf leaf NOUN fcis-7187 20 14 diseases disease NOUN fcis-7187 20 15 by by ADP fcis-7187 20 16 extracting extract VERB fcis-7187 20 17 image image NOUN fcis-7187 20 18 features feature VERB fcis-7187 20 19 automatically automatically ADV fcis-7187 20 20 . . PUNCT fcis-7187 21 1 gayathri gayathri PROPN fcis-7187 21 2 et et PROPN fcis-7187 21 3 al al PROPN fcis-7187 21 4 . . PUNCT fcis-7187 22 1 [ [ X fcis-7187 22 2 2 2 X fcis-7187 22 3 ] ] PUNCT fcis-7187 22 4 proposed propose VERB fcis-7187 22 5 a a DET fcis-7187 22 6 deep deep ADJ fcis-7187 22 7 cnns cnn NOUN fcis-7187 22 8 called call VERB fcis-7187 22 9 lenet lenet PROPN fcis-7187 22 10 to to PART fcis-7187 22 11 discover discover VERB fcis-7187 22 12 the the DET fcis-7187 22 13 tea tea NOUN fcis-7187 22 14 plant plant NOUN fcis-7187 22 15 diseases disease NOUN fcis-7187 22 16 from from ADP fcis-7187 22 17 leaf leaf NOUN fcis-7187 22 18 image image NOUN fcis-7187 22 19 set set NOUN fcis-7187 22 20 , , PUNCT fcis-7187 22 21 which which PRON fcis-7187 22 22 can can AUX fcis-7187 22 23 be be AUX fcis-7187 22 24 applied apply VERB fcis-7187 22 25 to to PART fcis-7187 22 26 improve improve VERB fcis-7187 22 27 the the DET fcis-7187 22 28 diagnostic diagnostic ADJ fcis-7187 22 29 measurement measurement NOUN fcis-7187 22 30 of of ADP fcis-7187 22 31 tea tea NOUN fcis-7187 22 32 leaves leave NOUN fcis-7187 22 33 . . PUNCT fcis-7187 23 1 hu hu PROPN fcis-7187 23 2 et et PROPN fcis-7187 23 3 al al PROPN fcis-7187 23 4 . . PUNCT fcis-7187 24 1 [ [ X fcis-7187 24 2 3 3 X fcis-7187 24 3 ] ] PUNCT fcis-7187 24 4 proposed propose VERB fcis-7187 24 5 an an DET fcis-7187 24 6 improved improved ADJ fcis-7187 24 7 deep deep ADJ fcis-7187 24 8 cnn cnn NOUN fcis-7187 24 9 for for ADP fcis-7187 24 10 tea tea NOUN fcis-7187 24 11 leaf leaf NOUN fcis-7187 24 12 disease disease NOUN fcis-7187 24 13 identification identification NOUN fcis-7187 24 14 , , PUNCT fcis-7187 24 15 in in ADP fcis-7187 24 16 which which PRON fcis-7187 24 17 the the DET fcis-7187 24 18 depthwise depthwise NOUN fcis-7187 24 19 separable separable ADJ fcis-7187 24 20 convolution convolution NOUN fcis-7187 24 21 is be AUX fcis-7187 24 22 used use VERB fcis-7187 24 23 to to PART fcis-7187 24 24 reduce reduce VERB fcis-7187 24 25 the the DET fcis-7187 24 26 number number NOUN fcis-7187 24 27 of of ADP fcis-7187 24 28 model model NOUN fcis-7187 24 29 parameters parameter NOUN fcis-7187 24 30 and and CCONJ fcis-7187 24 31 accelerate accelerate VERB fcis-7187 24 32 the the DET fcis-7187 24 33 calculation calculation NOUN fcis-7187 24 34 of of ADP fcis-7187 24 35 the the DET fcis-7187 24 36 model model NOUN fcis-7187 24 37 . . PUNCT fcis-7187 25 1 experiments experiment NOUN fcis-7187 25 2 show show VERB fcis-7187 25 3 that that SCONJ fcis-7187 25 4 the the DET fcis-7187 25 5 average average ADJ fcis-7187 25 6 identification identification NOUN fcis-7187 25 7 accuracy accuracy NOUN fcis-7187 25 8 is be AUX fcis-7187 25 9 higher high ADJ fcis-7187 25 10 than than ADP fcis-7187 25 11 classical classical ADJ fcis-7187 25 12 deep deep ADJ fcis-7187 25 13 learning learning NOUN fcis-7187 25 14 methods method NOUN fcis-7187 25 15 . . PUNCT fcis-7187 26 1 the the DET fcis-7187 26 2 effectiveness effectiveness NOUN fcis-7187 26 3 of of ADP fcis-7187 26 4 deep deep ADJ fcis-7187 26 5 learning learning NOUN fcis-7187 26 6 benefits benefit NOUN fcis-7187 26 7 from from ADP fcis-7187 26 8 the the DET fcis-7187 26 9 construction construction NOUN fcis-7187 26 10 of of ADP fcis-7187 26 11 deep deep ADJ fcis-7187 26 12 neural neural ADJ fcis-7187 26 13 networks network NOUN fcis-7187 26 14 , , PUNCT fcis-7187 26 15 and and CCONJ fcis-7187 26 16 its its PRON fcis-7187 26 17 training training NOUN fcis-7187 26 18 process process NOUN fcis-7187 26 19 requires require VERB fcis-7187 26 20 a a DET fcis-7187 26 21 large large ADJ fcis-7187 26 22 number number NOUN fcis-7187 26 23 of of ADP fcis-7187 26 24 parameters parameter NOUN fcis-7187 26 25 training training NOUN fcis-7187 26 26 tasks task NOUN fcis-7187 26 27 . . PUNCT fcis-7187 27 1 in in ADP fcis-7187 27 2 order order NOUN fcis-7187 27 3 to to PART fcis-7187 27 4 improve improve VERB fcis-7187 27 5 the the DET fcis-7187 27 6 generalization generalization NOUN fcis-7187 27 7 ability ability NOUN fcis-7187 27 8 of of ADP fcis-7187 27 9 the the DET fcis-7187 27 10 neural neural ADJ fcis-7187 27 11 network network NOUN fcis-7187 27 12 models model NOUN fcis-7187 27 13 and and CCONJ fcis-7187 27 14 avoid avoid VERB fcis-7187 27 15 the the DET fcis-7187 27 16 overfitting overfitting NOUN fcis-7187 27 17 problem problem NOUN fcis-7187 27 18 , , PUNCT fcis-7187 27 19 the the DET fcis-7187 27 20 model model NOUN fcis-7187 27 21 training training NOUN fcis-7187 27 22 process process NOUN fcis-7187 27 23 usually usually ADV fcis-7187 27 24 requires require VERB fcis-7187 27 25 the the DET fcis-7187 27 26 input input NOUN fcis-7187 27 27 of of ADP fcis-7187 27 28 a a DET fcis-7187 27 29 large large ADJ fcis-7187 27 30 number number NOUN fcis-7187 27 31 of of ADP fcis-7187 27 32 labeled label VERB fcis-7187 27 33 data data NOUN fcis-7187 27 34 samples sample NOUN fcis-7187 27 35 for for ADP fcis-7187 27 36 learning learn VERB fcis-7187 27 37 . . PUNCT fcis-7187 28 1 however however ADV fcis-7187 28 2 , , PUNCT fcis-7187 28 3 in in ADP fcis-7187 28 4 practical practical ADJ fcis-7187 28 5 application application NOUN fcis-7187 28 6 scenarios scenario NOUN fcis-7187 28 7 , , PUNCT fcis-7187 28 8 the the DET fcis-7187 28 9 acquisition acquisition NOUN fcis-7187 28 10 of of ADP fcis-7187 28 11 data datum NOUN fcis-7187 28 12 samples sample NOUN fcis-7187 28 13 may may AUX fcis-7187 28 14 have have VERB fcis-7187 28 15 environmental environmental ADJ fcis-7187 28 16 changes change NOUN fcis-7187 28 17 , , PUNCT fcis-7187 28 18 such such ADJ fcis-7187 28 19 as as ADP fcis-7187 28 20 the the DET fcis-7187 28 21 change change NOUN fcis-7187 28 22 of of ADP fcis-7187 28 23 illumination illumination NOUN fcis-7187 28 24 intensity intensity NOUN fcis-7187 28 25 , , PUNCT fcis-7187 28 26 shooting shoot VERB fcis-7187 28 27 equipment equipment NOUN fcis-7187 28 28 and and CCONJ fcis-7187 28 29 shooting shooting NOUN fcis-7187 28 30 angles angle NOUN fcis-7187 28 31 . . PUNCT fcis-7187 29 1 changes change NOUN fcis-7187 29 2 in in ADP fcis-7187 29 3 the the DET fcis-7187 29 4 environment environment NOUN fcis-7187 29 5 often often ADV fcis-7187 29 6 lead lead VERB fcis-7187 29 7 to to ADP fcis-7187 29 8 different different ADJ fcis-7187 29 9 data datum NOUN fcis-7187 29 10 distribution distribution NOUN fcis-7187 29 11 of of ADP fcis-7187 29 12 the the DET fcis-7187 29 13 acquired acquire VERB fcis-7187 29 14 sample sample NOUN fcis-7187 29 15 images image NOUN fcis-7187 29 16 , , PUNCT fcis-7187 29 17 thus thus ADV fcis-7187 29 18 reducing reduce VERB fcis-7187 29 19 the the DET fcis-7187 29 20 effect effect NOUN fcis-7187 29 21 of of ADP fcis-7187 29 22 image image NOUN fcis-7187 29 23 recognition recognition NOUN fcis-7187 29 24 . . PUNCT fcis-7187 30 1 in in ADP fcis-7187 30 2 other other ADJ fcis-7187 30 3 words word NOUN fcis-7187 30 4 , , PUNCT fcis-7187 30 5 the the DET fcis-7187 30 6 model model NOUN fcis-7187 30 7 with with ADP fcis-7187 30 8 good good ADJ fcis-7187 30 9 recognition recognition NOUN fcis-7187 30 10 effect effect NOUN fcis-7187 30 11 of of ADP fcis-7187 30 12 one one NUM fcis-7187 30 13 kind kind NOUN fcis-7187 30 14 of of ADP fcis-7187 30 15 data datum NOUN fcis-7187 30 16 distribution distribution NOUN fcis-7187 30 17 may may AUX fcis-7187 30 18 show show VERB fcis-7187 30 19 poor poor ADJ fcis-7187 30 20 performance performance NOUN fcis-7187 30 21 in in ADP fcis-7187 30 22 another another DET fcis-7187 30 23 kind kind NOUN fcis-7187 30 24 of of ADP fcis-7187 30 25 data datum NOUN fcis-7187 30 26 distribution distribution NOUN fcis-7187 30 27 . . PUNCT fcis-7187 31 1 to to PART fcis-7187 31 2 solve solve VERB fcis-7187 31 3 this this DET fcis-7187 31 4 problem problem NOUN fcis-7187 31 5 , , PUNCT fcis-7187 31 6 domain domain NOUN fcis-7187 31 7 - - PUNCT fcis-7187 31 8 adaptive adaptive ADJ fcis-7187 31 9 methods method NOUN fcis-7187 31 10 emerge emerge VERB fcis-7187 31 11 . . PUNCT fcis-7187 32 1 this this DET fcis-7187 32 2 method method NOUN fcis-7187 32 3 is be AUX fcis-7187 32 4 a a DET fcis-7187 32 5 transfer transfer NOUN fcis-7187 32 6 learning learning NOUN fcis-7187 32 7 method method NOUN fcis-7187 32 8 aiming aim VERB fcis-7187 32 9 to to PART fcis-7187 32 10 transfer transfer VERB fcis-7187 32 11 the the DET fcis-7187 32 12 classification classification NOUN fcis-7187 32 13 recognition recognition NOUN fcis-7187 32 14 ability ability NOUN fcis-7187 32 15 learned learn VERB fcis-7187 32 16 in in ADP fcis-7187 32 17 the the DET fcis-7187 32 18 source source NOUN fcis-7187 32 19 domain domain NOUN fcis-7187 32 20 to to ADP fcis-7187 32 21 the the DET fcis-7187 32 22 target target NOUN fcis-7187 32 23 domain domain NOUN fcis-7187 32 24 . . PUNCT fcis-7187 33 1 domain domain NOUN fcis-7187 33 2 adaptive adaptive ADJ fcis-7187 33 3 methods method NOUN fcis-7187 33 4 divide divide VERB fcis-7187 33 5 the the DET fcis-7187 33 6 training training NOUN fcis-7187 33 7 dataset dataset VERB fcis-7187 33 8 into into ADP fcis-7187 33 9 labeled label VERB fcis-7187 33 10 source source NOUN fcis-7187 33 11 domain domain NOUN fcis-7187 33 12 and and CCONJ fcis-7187 33 13 target target NOUN fcis-7187 33 14 domain domain NOUN fcis-7187 33 15 with with ADP fcis-7187 33 16 little little ADJ fcis-7187 33 17 or or CCONJ fcis-7187 33 18 no no PRON fcis-7187 33 19 labels label NOUN fcis-7187 33 20 depending depend VERB fcis-7187 33 21 on on ADP fcis-7187 33 22 the the DET fcis-7187 33 23 data datum NOUN fcis-7187 33 24 distribution distribution NOUN fcis-7187 33 25 . . PUNCT fcis-7187 34 1 when when SCONJ fcis-7187 34 2 the the DET fcis-7187 34 3 model model NOUN fcis-7187 34 4 has have VERB fcis-7187 34 5 the the DET fcis-7187 34 6 ability ability NOUN fcis-7187 34 7 to to PART fcis-7187 34 8 recognize recognize VERB fcis-7187 34 9 the the DET fcis-7187 34 10 source source NOUN fcis-7187 34 11 domain domain NOUN fcis-7187 34 12 data datum NOUN fcis-7187 34 13 , , PUNCT fcis-7187 34 14 this this DET fcis-7187 34 15 kind kind NOUN fcis-7187 34 16 of of ADP fcis-7187 34 17 ability ability NOUN fcis-7187 34 18 can can AUX fcis-7187 34 19 be be AUX fcis-7187 34 20 quickly quickly ADV fcis-7187 34 21 transferred transfer VERB fcis-7187 34 22 to to ADP fcis-7187 34 23 the the DET fcis-7187 34 24 target target NOUN fcis-7187 34 25 domain domain NOUN fcis-7187 34 26 through through ADP fcis-7187 34 27 the the DET fcis-7187 34 28 domain domain NOUN fcis-7187 34 29 adaptive adaptive ADJ fcis-7187 34 30 method method NOUN fcis-7187 34 31 . . PUNCT fcis-7187 35 1 the the DET fcis-7187 35 2 basic basic ADJ fcis-7187 35 3 idea idea NOUN fcis-7187 35 4 of of ADP fcis-7187 35 5 domain domain NOUN fcis-7187 35 6 adaptation adaptation NOUN fcis-7187 35 7 is be AUX fcis-7187 35 8 to to PART fcis-7187 35 9 map map VERB fcis-7187 35 10 the the DET fcis-7187 35 11 data data NOUN fcis-7187 35 12 sets set NOUN fcis-7187 35 13 from from ADP fcis-7187 35 14 different different ADJ fcis-7187 35 15 domains domain NOUN fcis-7187 35 16 to to ADP fcis-7187 35 17 a a DET fcis-7187 35 18 same same ADJ fcis-7187 35 19 feature feature NOUN fcis-7187 35 20 space space NOUN fcis-7187 35 21 and and CCONJ fcis-7187 35 22 make make VERB fcis-7187 35 23 their their PRON fcis-7187 35 24 feature feature NOUN fcis-7187 35 25 distribution distribution NOUN fcis-7187 35 26 as as ADV fcis-7187 35 27 close close ADJ fcis-7187 35 28 as as ADP fcis-7187 35 29 possible possible ADJ fcis-7187 35 30 , , PUNCT fcis-7187 35 31 so so SCONJ fcis-7187 35 32 as as SCONJ fcis-7187 35 33 to to PART fcis-7187 35 34 realize realize VERB fcis-7187 35 35 cross cross ADJ fcis-7187 35 36 - - ADJ fcis-7187 35 37 domain domain ADJ fcis-7187 35 38 recognition recognition NOUN fcis-7187 35 39 . . PUNCT fcis-7187 36 1 there there PRON fcis-7187 36 2 are be VERB fcis-7187 36 3 many many ADJ fcis-7187 36 4 ways way NOUN fcis-7187 36 5 to to PART fcis-7187 36 6 realize realize VERB fcis-7187 36 7 domain domain NOUN fcis-7187 36 8 adaptation adaptation NOUN fcis-7187 36 9 , , PUNCT fcis-7187 36 10 among among ADP fcis-7187 36 11 which which PRON fcis-7187 36 12 the the DET fcis-7187 36 13 domain domain NOUN fcis-7187 36 14 adversarial adversarial ADJ fcis-7187 36 15 based base VERB fcis-7187 36 16 neural neural ADJ fcis-7187 36 17 network network NOUN fcis-7187 36 18 dann dann NOUN fcis-7187 36 19 [ [ X fcis-7187 36 20 4 4 NUM fcis-7187 36 21 ] ] PUNCT fcis-7187 36 22 is be AUX fcis-7187 36 23 proposed propose VERB fcis-7187 36 24 . . PUNCT fcis-7187 37 1 its its PRON fcis-7187 37 2 adversarial adversarial ADJ fcis-7187 37 3 based base VERB fcis-7187 37 4 idea idea NOUN fcis-7187 37 5 has have AUX fcis-7187 37 6 become become VERB fcis-7187 37 7 one one NUM fcis-7187 37 8 of of ADP fcis-7187 37 9 the the DET fcis-7187 37 10 main main ADJ fcis-7187 37 11 methods method NOUN fcis-7187 37 12 to to PART fcis-7187 37 13 realize realize VERB fcis-7187 37 14 domain domain NOUN fcis-7187 37 15 alignment alignment NOUN fcis-7187 37 16 . . PUNCT fcis-7187 38 1 chen chen PROPN fcis-7187 38 2 et et PROPN fcis-7187 38 3 al al PROPN fcis-7187 38 4 . . PUNCT fcis-7187 39 1 [ [ X fcis-7187 39 2 5 5 NUM fcis-7187 39 3 ] ] PUNCT fcis-7187 39 4 solve solve VERB fcis-7187 39 5 the the DET fcis-7187 39 6 domain domain NOUN fcis-7187 39 7 shift shift NOUN fcis-7187 39 8 problem problem NOUN fcis-7187 39 9 at at ADP fcis-7187 39 10 two two NUM fcis-7187 39 11 levels level NOUN fcis-7187 39 12 : : PUNCT fcis-7187 39 13 image image NOUN fcis-7187 39 14 level level NOUN fcis-7187 39 15 and and CCONJ fcis-7187 39 16 instance instance NOUN fcis-7187 39 17 level level NOUN fcis-7187 39 18 , , PUNCT fcis-7187 39 19 and and CCONJ fcis-7187 39 20 construct construct VERB fcis-7187 39 21 domain domain NOUN fcis-7187 39 22 classifiers classifier NOUN fcis-7187 39 23 by by ADP fcis-7187 39 24 using use VERB fcis-7187 39 25 adversarial adversarial ADJ fcis-7187 39 26 training training NOUN fcis-7187 39 27 to to PART fcis-7187 39 28 realize realize VERB fcis-7187 39 29 cross cross ADJ fcis-7187 39 30 - - ADJ fcis-7187 39 31 domain domain ADJ fcis-7187 39 32 object object NOUN fcis-7187 39 33 detection detection NOUN fcis-7187 39 34 . . PUNCT fcis-7187 40 1 hsu hsu PROPN fcis-7187 40 2 et et PROPN fcis-7187 40 3 al al PROPN fcis-7187 40 4 . . PUNCT fcis-7187 41 1 [ [ X fcis-7187 41 2 6 6 NUM fcis-7187 41 3 ] ] PUNCT fcis-7187 41 4 realize realize VERB fcis-7187 41 5 the the DET fcis-7187 41 6 conversion conversion NOUN fcis-7187 41 7 of of ADP fcis-7187 41 8 the the DET fcis-7187 41 9 source source NOUN fcis-7187 41 10 domain domain NOUN fcis-7187 41 11 to to ADP fcis-7187 41 12 the the DET fcis-7187 41 13 target target NOUN fcis-7187 41 14 domain domain NOUN fcis-7187 41 15 by by ADP fcis-7187 41 16 constructing construct VERB fcis-7187 41 17 the the DET fcis-7187 41 18 transition transition NOUN fcis-7187 41 19 domain domain NOUN fcis-7187 41 20 , , PUNCT fcis-7187 41 21 and and CCONJ fcis-7187 41 22 adopts adopt VERB fcis-7187 41 23 the the DET fcis-7187 41 24 adversarial adversarial ADJ fcis-7187 41 25 training training NOUN fcis-7187 41 26 method method NOUN fcis-7187 41 27 to to PART fcis-7187 41 28 achieve achieve VERB fcis-7187 41 29 domain domain NOUN fcis-7187 41 30 alignment alignment NOUN fcis-7187 41 31 at at ADP fcis-7187 41 32 the the DET fcis-7187 41 33 feature feature NOUN fcis-7187 41 34 level level NOUN fcis-7187 41 35 . . PUNCT fcis-7187 42 1 tseng tseng PROPN fcis-7187 42 2 et et PROPN fcis-7187 42 3 al al PROPN fcis-7187 42 4 . . PUNCT fcis-7187 43 1 [ [ X fcis-7187 43 2 7 7 X fcis-7187 43 3 ] ] PUNCT fcis-7187 43 4 use use VERB fcis-7187 43 5 feature feature NOUN fcis-7187 43 6 - - PUNCT fcis-7187 43 7 wise wise ADJ fcis-7187 43 8 transformation transformation NOUN fcis-7187 43 9 layers layer NOUN fcis-7187 43 10 by by ADP fcis-7187 43 11 using use VERB fcis-7187 43 12 affine affine NOUN fcis-7187 43 13 transforms transform VERB fcis-7187 43 14 to to PART fcis-7187 43 15 simulate simulate VERB fcis-7187 43 16 various various ADJ fcis-7187 43 17 feature feature NOUN fcis-7187 43 18 distributions distribution NOUN fcis-7187 43 19 under under ADP fcis-7187 43 20 different different ADJ fcis-7187 43 21 domains domain NOUN fcis-7187 43 22 . . PUNCT fcis-7187 44 1 there there PRON fcis-7187 44 2 experiments experiment NOUN fcis-7187 44 3 show show VERB fcis-7187 44 4 improvements improvement NOUN fcis-7187 44 5 on on ADP fcis-7187 44 6 the the DET fcis-7187 44 7 few few ADJ fcis-7187 44 8 - - PUNCT fcis-7187 44 9 shot shot NOUN fcis-7187 44 10 classification classification NOUN fcis-7187 44 11 performance performance NOUN fcis-7187 44 12 under under ADP fcis-7187 44 13 domain domain NOUN fcis-7187 44 14 shift shift NOUN fcis-7187 44 15 . . PUNCT fcis-7187 45 1 2 2 X fcis-7187 45 2 . . X fcis-7187 45 3 related relate VERB fcis-7187 45 4 theory theory NOUN fcis-7187 45 5 2.1 2.1 NUM fcis-7187 45 6 . . PUNCT fcis-7187 46 1 convolutional convolutional ADJ fcis-7187 46 2 neural neural ADJ fcis-7187 46 3 network network NOUN fcis-7187 46 4 convolutional convolutional ADJ fcis-7187 46 5 neural neural ADJ fcis-7187 46 6 networks network NOUN fcis-7187 46 7 ( ( PUNCT fcis-7187 46 8 cnn cnn PROPN fcis-7187 46 9 ) ) PUNCT fcis-7187 46 10 is be AUX fcis-7187 46 11 one one NUM fcis-7187 46 12 of of ADP fcis-7187 46 13 the the DET fcis-7187 46 14 key key ADJ fcis-7187 46 15 factors factor NOUN fcis-7187 46 16 for for ADP fcis-7187 46 17 the the DET fcis-7187 46 18 success success NOUN fcis-7187 46 19 of of ADP fcis-7187 46 20 deep deep ADJ fcis-7187 46 21 learning learning NOUN fcis-7187 46 22 in in ADP fcis-7187 46 23 the the DET fcis-7187 46 24 field field NOUN fcis-7187 46 25 of of ADP fcis-7187 46 26 computer computer NOUN fcis-7187 46 27 vision vision NOUN fcis-7187 46 28 recognition recognition PROPN fcis-7187 46 29 . . PUNCT fcis-7187 47 1 unlike unlike ADP fcis-7187 47 2 traditional traditional ADJ fcis-7187 47 3 machine machine NOUN fcis-7187 47 4 learning learning NOUN fcis-7187 47 5 methods method NOUN fcis-7187 47 6 , , PUNCT fcis-7187 47 7 which which PRON fcis-7187 47 8 require require VERB fcis-7187 47 9 manual manual ADJ fcis-7187 47 10 feature feature NOUN fcis-7187 47 11 extraction extraction NOUN fcis-7187 47 12 , , PUNCT fcis-7187 47 13 cnn cnn PROPN fcis-7187 47 14 can can AUX fcis-7187 47 15 achieve achieve VERB fcis-7187 47 16 end end NOUN fcis-7187 47 17 - - PUNCT fcis-7187 47 18 to to ADP fcis-7187 47 19 - - PUNCT fcis-7187 47 20 end end VERB fcis-7187 47 21 automatic automatic ADJ fcis-7187 47 22 feature feature NOUN fcis-7187 47 23 extraction extraction NOUN fcis-7187 47 24 for for ADP fcis-7187 47 25 input input NOUN fcis-7187 47 26 images image NOUN fcis-7187 47 27 . . PUNCT fcis-7187 48 1 the the DET fcis-7187 48 2 typical typical ADJ fcis-7187 48 3 cnn cnn PROPN fcis-7187 48 4 structure structure NOUN fcis-7187 48 5 is be AUX fcis-7187 48 6 mainly mainly ADV fcis-7187 48 7 composed compose VERB fcis-7187 48 8 49 49 NUM fcis-7187 48 9 of of ADP fcis-7187 48 10 convolutional convolutional ADJ fcis-7187 48 11 layers layer NOUN fcis-7187 48 12 , , PUNCT fcis-7187 48 13 pooling pool VERB fcis-7187 48 14 layers layer NOUN fcis-7187 48 15 and and CCONJ fcis-7187 48 16 fully fully ADV fcis-7187 48 17 connected connected ADJ fcis-7187 48 18 layers layer NOUN fcis-7187 48 19 . . PUNCT fcis-7187 49 1 convolutional convolutional ADJ fcis-7187 49 2 layer layer NOUN fcis-7187 49 3 is be AUX fcis-7187 49 4 one one NUM fcis-7187 49 5 of of ADP fcis-7187 49 6 the the DET fcis-7187 49 7 important important ADJ fcis-7187 49 8 structures structure NOUN fcis-7187 49 9 of of ADP fcis-7187 49 10 cnn cnn PROPN fcis-7187 49 11 . . PUNCT fcis-7187 50 1 it it PRON fcis-7187 50 2 performs perform VERB fcis-7187 50 3 convolution convolution NOUN fcis-7187 50 4 computation computation NOUN fcis-7187 50 5 by by ADP fcis-7187 50 6 convolutional convolutional ADJ fcis-7187 50 7 kernels kernel NOUN fcis-7187 50 8 for for ADP fcis-7187 50 9 the the DET fcis-7187 50 10 input input NOUN fcis-7187 50 11 features feature NOUN fcis-7187 50 12 , , PUNCT fcis-7187 50 13 so so SCONJ fcis-7187 50 14 as as SCONJ fcis-7187 50 15 to to PART fcis-7187 50 16 realize realize VERB fcis-7187 50 17 feature feature NOUN fcis-7187 50 18 extraction extraction NOUN fcis-7187 50 19 . . PUNCT fcis-7187 51 1 the the DET fcis-7187 51 2 model model NOUN fcis-7187 51 3 parameters parameter NOUN fcis-7187 51 4 of of ADP fcis-7187 51 5 the the DET fcis-7187 51 6 convolutional convolutional ADJ fcis-7187 51 7 layer layer NOUN fcis-7187 51 8 are be AUX fcis-7187 51 9 divided divide VERB fcis-7187 51 10 into into ADP fcis-7187 51 11 the the DET fcis-7187 51 12 convolutional convolutional ADJ fcis-7187 51 13 kernel kernel NOUN fcis-7187 51 14 part part NOUN fcis-7187 51 15 and and CCONJ fcis-7187 51 16 the the DET fcis-7187 51 17 deviation deviation NOUN fcis-7187 51 18 part part NOUN fcis-7187 51 19 . . PUNCT fcis-7187 52 1 a a DET fcis-7187 52 2 convolutional convolutional ADJ fcis-7187 52 3 network network NOUN fcis-7187 52 4 with with ADP fcis-7187 52 5 different different ADJ fcis-7187 52 6 structure structure NOUN fcis-7187 52 7 and and CCONJ fcis-7187 52 8 complexity complexity NOUN fcis-7187 52 9 can can AUX fcis-7187 52 10 be be AUX fcis-7187 52 11 designed design VERB fcis-7187 52 12 between between ADP fcis-7187 52 13 multiple multiple ADJ fcis-7187 52 14 convolutional convolutional ADJ fcis-7187 52 15 layers layer NOUN fcis-7187 52 16 through through ADP fcis-7187 52 17 parallel parallel ADJ fcis-7187 52 18 or or CCONJ fcis-7187 52 19 serial serial ADJ fcis-7187 52 20 connection connection NOUN fcis-7187 52 21 . . PUNCT fcis-7187 53 1 the the DET fcis-7187 53 2 characteristic characteristic ADJ fcis-7187 53 3 output output NOUN fcis-7187 53 4 of of ADP fcis-7187 53 5 the the DET fcis-7187 53 6 i i PROPN fcis-7187 53 7 - - PUNCT fcis-7187 53 8 th th VERB fcis-7187 53 9 layer layer NOUN fcis-7187 53 10 in in ADP fcis-7187 53 11 the the DET fcis-7187 53 12 convolutional convolutional ADJ fcis-7187 53 13 network network NOUN fcis-7187 53 14 can can AUX fcis-7187 53 15 be be AUX fcis-7187 53 16 expressed express VERB fcis-7187 53 17 as as ADP fcis-7187 53 18 : : PUNCT fcis-7187 53 19 𝑦𝑖 𝑦𝑖 PROPN fcis-7187 53 20 = = PUNCT fcis-7187 53 21 𝑓(π‘¦π‘–βˆ’1π‘Šπ‘– 𝑓(π‘¦π‘–βˆ’1π‘Šπ‘– PROPN fcis-7187 53 22 + + CCONJ fcis-7187 53 23 𝑏𝑖 𝑏𝑖 NOUN fcis-7187 53 24 ) ) PUNCT fcis-7187 53 25 where where SCONJ fcis-7187 53 26 π‘Šπ‘– π‘Šπ‘– PROPN fcis-7187 53 27 is be AUX fcis-7187 53 28 the the DET fcis-7187 53 29 weight weight NOUN fcis-7187 53 30 parameter parameter NOUN fcis-7187 53 31 of of ADP fcis-7187 53 32 the the DET fcis-7187 53 33 convolution convolution NOUN fcis-7187 53 34 kernel kernel NOUN fcis-7187 53 35 in in ADP fcis-7187 53 36 i i PROPN fcis-7187 53 37 - - PUNCT fcis-7187 53 38 th th X fcis-7187 53 39 layer layer NOUN fcis-7187 53 40 , , PUNCT fcis-7187 53 41 𝑏𝑖 𝑏𝑖 PROPN fcis-7187 53 42 is be AUX fcis-7187 53 43 the the DET fcis-7187 53 44 standard standard ADJ fcis-7187 53 45 deviation deviation NOUN fcis-7187 53 46 value value NOUN fcis-7187 53 47 , , PUNCT fcis-7187 53 48 π‘¦π‘–βˆ’1 π‘¦π‘–βˆ’1 NOUN fcis-7187 53 49 is be AUX fcis-7187 53 50 the the DET fcis-7187 53 51 feature feature NOUN fcis-7187 53 52 input input NOUN fcis-7187 53 53 of of ADP fcis-7187 53 54 the the DET fcis-7187 53 55 convolutional convolutional ADJ fcis-7187 53 56 layer layer NOUN fcis-7187 53 57 ( ( PUNCT fcis-7187 53 58 𝑦0 𝑦0 NOUN fcis-7187 53 59 represents represent VERB fcis-7187 53 60 the the DET fcis-7187 53 61 input input NOUN fcis-7187 53 62 image image NOUN fcis-7187 53 63 ) ) PUNCT fcis-7187 53 64 , , PUNCT fcis-7187 53 65 and and CCONJ fcis-7187 53 66 f f PROPN fcis-7187 53 67 represents represent VERB fcis-7187 53 68 the the DET fcis-7187 53 69 activation activation NOUN fcis-7187 53 70 function function NOUN fcis-7187 53 71 . . PUNCT fcis-7187 54 1 the the DET fcis-7187 54 2 main main ADJ fcis-7187 54 3 function function NOUN fcis-7187 54 4 of of ADP fcis-7187 54 5 the the DET fcis-7187 54 6 pooling pooling NOUN fcis-7187 54 7 layer layer NOUN fcis-7187 54 8 is be AUX fcis-7187 54 9 to to PART fcis-7187 54 10 strengthen strengthen VERB fcis-7187 54 11 the the DET fcis-7187 54 12 spatial spatial ADJ fcis-7187 54 13 invariance invariance NOUN fcis-7187 54 14 of of ADP fcis-7187 54 15 the the DET fcis-7187 54 16 features feature NOUN fcis-7187 54 17 extracted extract VERB fcis-7187 54 18 by by ADP fcis-7187 54 19 the the DET fcis-7187 54 20 convolution convolution NOUN fcis-7187 54 21 layers layer NOUN fcis-7187 54 22 and and CCONJ fcis-7187 54 23 reduce reduce VERB fcis-7187 54 24 the the DET fcis-7187 54 25 excessive excessive ADJ fcis-7187 54 26 sensitivity sensitivity NOUN fcis-7187 54 27 to to ADP fcis-7187 54 28 the the DET fcis-7187 54 29 spatial spatial ADJ fcis-7187 54 30 position position NOUN fcis-7187 54 31 , , PUNCT fcis-7187 54 32 so so SCONJ fcis-7187 54 33 that that SCONJ fcis-7187 54 34 the the DET fcis-7187 54 35 same same ADJ fcis-7187 54 36 object object NOUN fcis-7187 54 37 can can AUX fcis-7187 54 38 be be AUX fcis-7187 54 39 well well ADV fcis-7187 54 40 recognized recognize VERB fcis-7187 54 41 by by ADP fcis-7187 54 42 the the DET fcis-7187 54 43 model model NOUN fcis-7187 54 44 even even ADV fcis-7187 54 45 if if SCONJ fcis-7187 54 46 there there PRON fcis-7187 54 47 are be VERB fcis-7187 54 48 position position NOUN fcis-7187 54 49 changes change NOUN fcis-7187 54 50 . . PUNCT fcis-7187 55 1 the the DET fcis-7187 55 2 common common ADJ fcis-7187 55 3 pooling pooling NOUN fcis-7187 55 4 methods method NOUN fcis-7187 55 5 include include VERB fcis-7187 55 6 maximum maximum ADJ fcis-7187 55 7 pooling pooling NOUN fcis-7187 55 8 and and CCONJ fcis-7187 55 9 average average ADJ fcis-7187 55 10 pooling pooling NOUN fcis-7187 55 11 , , PUNCT fcis-7187 55 12 which which PRON fcis-7187 55 13 calculate calculate VERB fcis-7187 55 14 the the DET fcis-7187 55 15 maximum maximum ADJ fcis-7187 55 16 value value NOUN fcis-7187 55 17 and and CCONJ fcis-7187 55 18 average average ADJ fcis-7187 55 19 value value NOUN fcis-7187 55 20 in in ADP fcis-7187 55 21 the the DET fcis-7187 55 22 pooling pooling NOUN fcis-7187 55 23 window window NOUN fcis-7187 55 24 respectively respectively ADV fcis-7187 55 25 , , PUNCT fcis-7187 55 26 so so SCONJ fcis-7187 55 27 as as SCONJ fcis-7187 55 28 to to PART fcis-7187 55 29 realize realize VERB fcis-7187 55 30 the the DET fcis-7187 55 31 down down ADP fcis-7187 55 32 sampling sample VERB fcis-7187 55 33 process process NOUN fcis-7187 55 34 of of ADP fcis-7187 55 35 features feature NOUN fcis-7187 55 36 . . PUNCT fcis-7187 56 1 the the DET fcis-7187 56 2 characteristic characteristic ADJ fcis-7187 56 3 output output NOUN fcis-7187 56 4 of of ADP fcis-7187 56 5 the the DET fcis-7187 56 6 j j PROPN fcis-7187 56 7 - - PUNCT fcis-7187 56 8 th th VERB fcis-7187 56 9 pool pool NOUN fcis-7187 56 10 region region NOUN fcis-7187 56 11 of of ADP fcis-7187 56 12 the the DET fcis-7187 56 13 s s NOUN fcis-7187 56 14 - - PUNCT fcis-7187 56 15 th th VERB fcis-7187 56 16 pool pool NOUN fcis-7187 56 17 layer layer NOUN fcis-7187 56 18 can can AUX fcis-7187 56 19 be be AUX fcis-7187 56 20 expressed express VERB fcis-7187 56 21 as as ADP fcis-7187 56 22 : : PUNCT fcis-7187 56 23 x𝑗 x𝑗 NOUN fcis-7187 56 24 𝑠 𝑠 NOUN fcis-7187 56 25 = = PUNCT fcis-7187 57 1 𝑑(x𝑗 𝑑(x𝑗 PROPN fcis-7187 57 2 π‘ βˆ’1 π‘ βˆ’1 PROPN fcis-7187 57 3 , , PUNCT fcis-7187 57 4 𝑝 𝑝 PROPN fcis-7187 57 5 , , PUNCT fcis-7187 57 6 π‘ž π‘ž NOUN fcis-7187 57 7 ) ) PUNCT fcis-7187 57 8 where where SCONJ fcis-7187 57 9 x𝑗 x𝑗 PROPN fcis-7187 57 10 π‘ βˆ’1 π‘ βˆ’1 PROPN fcis-7187 57 11 represents represent VERB fcis-7187 57 12 the the DET fcis-7187 57 13 feature feature NOUN fcis-7187 57 14 input input NOUN fcis-7187 57 15 for for ADP fcis-7187 57 16 this this DET fcis-7187 57 17 pooling pool VERB fcis-7187 57 18 layer layer NOUN fcis-7187 57 19 , , PUNCT fcis-7187 57 20 p p NOUN fcis-7187 57 21 and and CCONJ fcis-7187 57 22 q q AUX fcis-7187 57 23 correspond correspond NOUN fcis-7187 57 24 to to ADP fcis-7187 57 25 the the DET fcis-7187 57 26 height height NOUN fcis-7187 57 27 and and CCONJ fcis-7187 57 28 width width NOUN fcis-7187 57 29 of of ADP fcis-7187 57 30 the the DET fcis-7187 57 31 pooling pool VERB fcis-7187 57 32 window window NOUN fcis-7187 57 33 respectively respectively ADV fcis-7187 57 34 , , PUNCT fcis-7187 57 35 and and CCONJ fcis-7187 57 36 d d NOUN fcis-7187 57 37 represents represent VERB fcis-7187 57 38 the the DET fcis-7187 57 39 pooling pool VERB fcis-7187 57 40 function function NOUN fcis-7187 57 41 . . PUNCT fcis-7187 58 1 after after ADP fcis-7187 58 2 a a DET fcis-7187 58 3 series series NOUN fcis-7187 58 4 of of ADP fcis-7187 58 5 convolution convolution NOUN fcis-7187 58 6 and and CCONJ fcis-7187 58 7 pooling pool VERB fcis-7187 58 8 operations operation NOUN fcis-7187 58 9 , , PUNCT fcis-7187 58 10 one one NUM fcis-7187 58 11 or or CCONJ fcis-7187 58 12 more more ADV fcis-7187 58 13 fully fully ADV fcis-7187 58 14 connected connected ADJ fcis-7187 58 15 layers layer NOUN fcis-7187 58 16 are be AUX fcis-7187 58 17 usually usually ADV fcis-7187 58 18 followed follow VERB fcis-7187 58 19 for for ADP fcis-7187 58 20 the the DET fcis-7187 58 21 image image NOUN fcis-7187 58 22 classification classification NOUN fcis-7187 58 23 task task NOUN fcis-7187 58 24 . . PUNCT fcis-7187 59 1 softmax softmax PROPN fcis-7187 59 2 function function NOUN fcis-7187 59 3 is be AUX fcis-7187 59 4 usually usually ADV fcis-7187 59 5 used use VERB fcis-7187 59 6 to to PART fcis-7187 59 7 map map VERB fcis-7187 59 8 the the DET fcis-7187 59 9 output output NOUN fcis-7187 59 10 of of ADP fcis-7187 59 11 the the DET fcis-7187 59 12 final final ADJ fcis-7187 59 13 network network NOUN fcis-7187 59 14 layer layer NOUN fcis-7187 59 15 into into ADP fcis-7187 59 16 a a DET fcis-7187 59 17 probability probability NOUN fcis-7187 59 18 distribution distribution NOUN fcis-7187 59 19 , , PUNCT fcis-7187 59 20 in in ADP fcis-7187 59 21 which which PRON fcis-7187 59 22 all all DET fcis-7187 59 23 values value NOUN fcis-7187 59 24 are be AUX fcis-7187 59 25 positive positive ADJ fcis-7187 59 26 and and CCONJ fcis-7187 59 27 the the DET fcis-7187 59 28 sum sum NOUN fcis-7187 59 29 is be AUX fcis-7187 59 30 1 1 NUM fcis-7187 59 31 . . PUNCT fcis-7187 60 1 the the DET fcis-7187 60 2 j j PROPN fcis-7187 60 3 - - PUNCT fcis-7187 60 4 th th VERB fcis-7187 60 5 output output NOUN fcis-7187 60 6 result result NOUN fcis-7187 60 7 can can AUX fcis-7187 60 8 be be AUX fcis-7187 60 9 expressed express VERB fcis-7187 60 10 as as ADP fcis-7187 60 11 : : PUNCT fcis-7187 60 12 𝑂𝑗 𝑂𝑗 PROPN fcis-7187 60 13 = = PUNCT fcis-7187 60 14 π‘†π‘œπ‘“π‘‘π‘šπ‘Žπ‘₯(𝑧𝑗 π‘†π‘œπ‘“π‘‘π‘šπ‘Žπ‘₯(𝑧𝑗 PROPN fcis-7187 60 15 ) ) PUNCT fcis-7187 61 1 = = SYM fcis-7187 61 2 𝑒𝑧𝑗/βˆ‘ 𝑒𝑧𝑗/βˆ‘ ADJ fcis-7187 61 3 π‘’π‘§π‘˜ π‘’π‘§π‘˜ NOUN fcis-7187 62 1 𝐾 𝐾 PROPN fcis-7187 63 1 π‘˜=1 π‘˜=1 PROPN fcis-7187 64 1 where where SCONJ fcis-7187 64 2 z z NOUN fcis-7187 64 3 represents represent VERB fcis-7187 64 4 the the DET fcis-7187 64 5 output output NOUN fcis-7187 64 6 vector vector NOUN fcis-7187 64 7 of of ADP fcis-7187 64 8 the the DET fcis-7187 64 9 last last ADJ fcis-7187 64 10 connected connected ADJ fcis-7187 64 11 layer layer NOUN fcis-7187 64 12 . . PUNCT fcis-7187 65 1 k k PROPN fcis-7187 65 2 is be AUX fcis-7187 65 3 the the DET fcis-7187 65 4 output output NOUN fcis-7187 65 5 numbers number NOUN fcis-7187 65 6 of of ADP fcis-7187 65 7 the the DET fcis-7187 65 8 last last ADJ fcis-7187 65 9 connected connected ADJ fcis-7187 65 10 layer layer NOUN fcis-7187 65 11 , , PUNCT fcis-7187 65 12 and and CCONJ fcis-7187 65 13 j∈ j∈ PROPN fcis-7187 65 14 { { PUNCT fcis-7187 65 15 1 1 NUM fcis-7187 65 16 , , PUNCT fcis-7187 65 17 … … PUNCT fcis-7187 65 18 , , PUNCT fcis-7187 65 19 k k NOUN fcis-7187 65 20 } } PUNCT fcis-7187 65 21 . . PUNCT fcis-7187 66 1 2.2 2.2 NUM fcis-7187 66 2 . . PUNCT fcis-7187 67 1 domain domain NOUN fcis-7187 67 2 adaptation adaptation NOUN fcis-7187 67 3 the the DET fcis-7187 67 4 traditional traditional ADJ fcis-7187 67 5 deep deep ADJ fcis-7187 67 6 learning learning NOUN fcis-7187 67 7 model model NOUN fcis-7187 67 8 is be AUX fcis-7187 67 9 mainly mainly ADV fcis-7187 67 10 trained train VERB fcis-7187 67 11 on on ADP fcis-7187 67 12 the the DET fcis-7187 67 13 training training NOUN fcis-7187 67 14 samples sample NOUN fcis-7187 67 15 with with ADP fcis-7187 67 16 the the DET fcis-7187 67 17 same same ADJ fcis-7187 67 18 data datum NOUN fcis-7187 67 19 distribution distribution NOUN fcis-7187 67 20 , , PUNCT fcis-7187 67 21 and and CCONJ fcis-7187 67 22 is be AUX fcis-7187 67 23 applied apply VERB fcis-7187 67 24 to to ADP fcis-7187 67 25 the the DET fcis-7187 67 26 test test NOUN fcis-7187 67 27 samples sample NOUN fcis-7187 67 28 with with ADP fcis-7187 67 29 the the DET fcis-7187 67 30 same same ADJ fcis-7187 67 31 distribution distribution NOUN fcis-7187 67 32 . . PUNCT fcis-7187 68 1 when when SCONJ fcis-7187 68 2 there there PRON fcis-7187 68 3 are be VERB fcis-7187 68 4 different different ADJ fcis-7187 68 5 data datum NOUN fcis-7187 68 6 distributions distribution NOUN fcis-7187 68 7 between between ADP fcis-7187 68 8 training training NOUN fcis-7187 68 9 samples sample NOUN fcis-7187 68 10 and and CCONJ fcis-7187 68 11 test test NOUN fcis-7187 68 12 samples sample NOUN fcis-7187 68 13 , , PUNCT fcis-7187 68 14 the the DET fcis-7187 68 15 learned learn VERB fcis-7187 68 16 model model NOUN fcis-7187 68 17 based base VERB fcis-7187 68 18 on on ADP fcis-7187 68 19 a a DET fcis-7187 68 20 certain certain ADJ fcis-7187 68 21 data data NOUN fcis-7187 68 22 distribution distribution NOUN fcis-7187 68 23 usually usually ADV fcis-7187 68 24 can can AUX fcis-7187 68 25 not not PART fcis-7187 68 26 adapt adapt VERB fcis-7187 68 27 to to ADP fcis-7187 68 28 this this DET fcis-7187 68 29 difference difference NOUN fcis-7187 68 30 well well ADV fcis-7187 68 31 in in ADP fcis-7187 68 32 another another DET fcis-7187 68 33 data datum NOUN fcis-7187 68 34 distribution distribution NOUN fcis-7187 68 35 , , PUNCT fcis-7187 68 36 and and CCONJ fcis-7187 68 37 shows show VERB fcis-7187 68 38 poor poor ADJ fcis-7187 68 39 generalization generalization NOUN fcis-7187 68 40 in in ADP fcis-7187 68 41 the the DET fcis-7187 68 42 specific specific ADJ fcis-7187 68 43 classification classification NOUN fcis-7187 68 44 task task NOUN fcis-7187 68 45 . . PUNCT fcis-7187 69 1 domain domain NOUN fcis-7187 69 2 adaptation adaptation NOUN fcis-7187 69 3 is be AUX fcis-7187 69 4 the the DET fcis-7187 69 5 way way NOUN fcis-7187 69 6 to to PART fcis-7187 69 7 solve solve VERB fcis-7187 69 8 this this DET fcis-7187 69 9 problem problem NOUN fcis-7187 69 10 of of ADP fcis-7187 69 11 differences difference NOUN fcis-7187 69 12 in in ADP fcis-7187 69 13 data datum NOUN fcis-7187 69 14 distribution distribution NOUN fcis-7187 69 15 . . PUNCT fcis-7187 70 1 it it PRON fcis-7187 70 2 divides divide VERB fcis-7187 70 3 the the DET fcis-7187 70 4 data data NOUN fcis-7187 70 5 sets set NOUN fcis-7187 70 6 into into ADP fcis-7187 70 7 source source NOUN fcis-7187 70 8 domain domain NOUN fcis-7187 70 9 and and CCONJ fcis-7187 70 10 target target NOUN fcis-7187 70 11 domain domain NOUN fcis-7187 70 12 , , PUNCT fcis-7187 70 13 aiming aim VERB fcis-7187 70 14 to to PART fcis-7187 70 15 realize realize VERB fcis-7187 70 16 the the DET fcis-7187 70 17 data datum NOUN fcis-7187 70 18 alignment alignment NOUN fcis-7187 70 19 of of ADP fcis-7187 70 20 different different ADJ fcis-7187 70 21 domains domain NOUN fcis-7187 70 22 , , PUNCT fcis-7187 70 23 so so SCONJ fcis-7187 70 24 as as SCONJ fcis-7187 70 25 to to PART fcis-7187 70 26 transfer transfer VERB fcis-7187 70 27 the the DET fcis-7187 70 28 recognition recognition NOUN fcis-7187 70 29 ability ability NOUN fcis-7187 70 30 of of ADP fcis-7187 70 31 the the DET fcis-7187 70 32 source source NOUN fcis-7187 70 33 domain domain NOUN fcis-7187 70 34 to to ADP fcis-7187 70 35 the the DET fcis-7187 70 36 target target NOUN fcis-7187 70 37 domain domain NOUN fcis-7187 70 38 . . PUNCT fcis-7187 71 1 therefore therefore ADV fcis-7187 71 2 , , PUNCT fcis-7187 71 3 the the DET fcis-7187 71 4 domain domain NOUN fcis-7187 71 5 adaptive adaptive ADJ fcis-7187 71 6 method method NOUN fcis-7187 71 7 is be AUX fcis-7187 71 8 essentially essentially ADV fcis-7187 71 9 a a DET fcis-7187 71 10 transfer transfer NOUN fcis-7187 71 11 learning learning NOUN fcis-7187 71 12 method method NOUN fcis-7187 71 13 . . PUNCT fcis-7187 72 1 3 3 X fcis-7187 72 2 . . X fcis-7187 72 3 method method NOUN fcis-7187 72 4 this this DET fcis-7187 72 5 study study NOUN fcis-7187 72 6 proposes propose VERB fcis-7187 72 7 a a DET fcis-7187 72 8 tea tea NOUN fcis-7187 72 9 leaf leaf NOUN fcis-7187 72 10 disease disease NOUN fcis-7187 72 11 classification classification NOUN fcis-7187 72 12 method method NOUN fcis-7187 72 13 based base VERB fcis-7187 72 14 on on ADP fcis-7187 72 15 domain domain NOUN fcis-7187 72 16 adaptation adaptation NOUN fcis-7187 72 17 , , PUNCT fcis-7187 72 18 which which PRON fcis-7187 72 19 is be AUX fcis-7187 72 20 introduced introduce VERB fcis-7187 72 21 from from ADP fcis-7187 72 22 the the DET fcis-7187 72 23 aspects aspect NOUN fcis-7187 72 24 of of ADP fcis-7187 72 25 data datum NOUN fcis-7187 72 26 sets set NOUN fcis-7187 72 27 and and CCONJ fcis-7187 72 28 domain domain VERB fcis-7187 72 29 adaptive adaptive ADJ fcis-7187 72 30 method method NOUN fcis-7187 72 31 respectively respectively ADV fcis-7187 72 32 . . PUNCT fcis-7187 73 1 3.1 3.1 NUM fcis-7187 73 2 . . PUNCT fcis-7187 73 3 data datum NOUN fcis-7187 73 4 sets set VERB fcis-7187 73 5 the the DET fcis-7187 73 6 data data NOUN fcis-7187 73 7 sets set NOUN fcis-7187 73 8 of of ADP fcis-7187 73 9 tea tea NOUN fcis-7187 73 10 leaf leaf NOUN fcis-7187 73 11 diseases disease NOUN fcis-7187 73 12 selected select VERB fcis-7187 73 13 in in ADP fcis-7187 73 14 this this DET fcis-7187 73 15 study study NOUN fcis-7187 73 16 were be AUX fcis-7187 73 17 all all PRON fcis-7187 73 18 collected collect VERB fcis-7187 73 19 in in ADP fcis-7187 73 20 the the DET fcis-7187 73 21 tea tea NOUN fcis-7187 73 22 garden garden NOUN fcis-7187 73 23 in in ADP fcis-7187 73 24 tai'an tai'an PROPN fcis-7187 73 25 , , PUNCT fcis-7187 73 26 shandong shandong PROPN fcis-7187 73 27 province province PROPN fcis-7187 73 28 . . PUNCT fcis-7187 74 1 there there PRON fcis-7187 74 2 are be VERB fcis-7187 74 3 totally totally ADV fcis-7187 74 4 746 746 NUM fcis-7187 74 5 images image NOUN fcis-7187 74 6 including include VERB fcis-7187 74 7 healthy healthy ADJ fcis-7187 74 8 leaves leave NOUN fcis-7187 74 9 and and CCONJ fcis-7187 74 10 three three NUM fcis-7187 74 11 kinds kind NOUN fcis-7187 74 12 of of ADP fcis-7187 74 13 infected infected ADJ fcis-7187 74 14 leaves leave NOUN fcis-7187 74 15 , , PUNCT fcis-7187 74 16 namely namely ADV fcis-7187 74 17 tea tea NOUN fcis-7187 74 18 white white PROPN fcis-7187 74 19 star star PROPN fcis-7187 74 20 , , PUNCT fcis-7187 74 21 tea tea NOUN fcis-7187 74 22 leaf leaf NOUN fcis-7187 74 23 blight blight NOUN fcis-7187 74 24 , , PUNCT fcis-7187 74 25 and and CCONJ fcis-7187 74 26 tea tea NOUN fcis-7187 74 27 wheel wheel NOUN fcis-7187 74 28 spot spot NOUN fcis-7187 74 29 . . PUNCT fcis-7187 75 1 after after ADP fcis-7187 75 2 applying apply VERB fcis-7187 75 3 data datum NOUN fcis-7187 75 4 enhancement enhancement NOUN fcis-7187 75 5 technology technology NOUN fcis-7187 75 6 such such ADJ fcis-7187 75 7 as as ADP fcis-7187 75 8 random random ADJ fcis-7187 75 9 rotation rotation NOUN fcis-7187 75 10 , , PUNCT fcis-7187 75 11 random random ADJ fcis-7187 75 12 cropping cropping NOUN fcis-7187 75 13 , , PUNCT fcis-7187 75 14 random random ADJ fcis-7187 75 15 color color NOUN fcis-7187 75 16 change change NOUN fcis-7187 75 17 , , PUNCT fcis-7187 75 18 the the DET fcis-7187 75 19 number number NOUN fcis-7187 75 20 of of ADP fcis-7187 75 21 the the DET fcis-7187 75 22 collected collected ADJ fcis-7187 75 23 tea tea NOUN fcis-7187 75 24 leaf leaf NOUN fcis-7187 75 25 images image NOUN fcis-7187 75 26 were be AUX fcis-7187 75 27 expanded expand VERB fcis-7187 75 28 to to ADP fcis-7187 75 29 4610 4610 NUM fcis-7187 75 30 . . PUNCT fcis-7187 76 1 for for ADP fcis-7187 76 2 each each DET fcis-7187 76 3 category category NOUN fcis-7187 76 4 , , PUNCT fcis-7187 76 5 70 70 NUM fcis-7187 76 6 % % NOUN fcis-7187 76 7 of of ADP fcis-7187 76 8 the the DET fcis-7187 76 9 data datum NOUN fcis-7187 76 10 were be AUX fcis-7187 76 11 randomly randomly ADV fcis-7187 76 12 selected select VERB fcis-7187 76 13 as as ADP fcis-7187 76 14 the the DET fcis-7187 76 15 training training NOUN fcis-7187 76 16 set set NOUN fcis-7187 76 17 and and CCONJ fcis-7187 76 18 30 30 NUM fcis-7187 76 19 % % NOUN fcis-7187 76 20 as as ADP fcis-7187 76 21 the the DET fcis-7187 76 22 test test NOUN fcis-7187 76 23 set set NOUN fcis-7187 76 24 . . PUNCT fcis-7187 77 1 the the DET fcis-7187 77 2 data datum NOUN fcis-7187 77 3 of of ADP fcis-7187 77 4 each each DET fcis-7187 77 5 category category NOUN fcis-7187 77 6 in in ADP fcis-7187 77 7 the the DET fcis-7187 77 8 training training NOUN fcis-7187 77 9 set set NOUN fcis-7187 77 10 are be AUX fcis-7187 77 11 divided divide VERB fcis-7187 77 12 into into ADP fcis-7187 77 13 source source NOUN fcis-7187 77 14 domain domain NOUN fcis-7187 77 15 and and CCONJ fcis-7187 77 16 target target NOUN fcis-7187 77 17 domain domain NOUN fcis-7187 77 18 in in ADP fcis-7187 77 19 a a DET fcis-7187 77 20 ratio ratio NOUN fcis-7187 77 21 of of ADP fcis-7187 77 22 4:1 4:1 NUM fcis-7187 77 23 . . PUNCT fcis-7187 78 1 to to PART fcis-7187 78 2 simulate simulate VERB fcis-7187 78 3 the the DET fcis-7187 78 4 different different ADJ fcis-7187 78 5 data datum NOUN fcis-7187 78 6 distributions distribution NOUN fcis-7187 78 7 of of ADP fcis-7187 78 8 the the DET fcis-7187 78 9 source source NOUN fcis-7187 78 10 and and CCONJ fcis-7187 78 11 target target NOUN fcis-7187 78 12 domains domain NOUN fcis-7187 78 13 , , PUNCT fcis-7187 78 14 the the DET fcis-7187 78 15 brightness brightness NOUN fcis-7187 78 16 of of ADP fcis-7187 78 17 the the DET fcis-7187 78 18 target target NOUN fcis-7187 78 19 domain domain NOUN fcis-7187 78 20 data datum NOUN fcis-7187 78 21 and and CCONJ fcis-7187 78 22 the the DET fcis-7187 78 23 raw raw ADJ fcis-7187 78 24 images image NOUN fcis-7187 78 25 in in ADP fcis-7187 78 26 the the DET fcis-7187 78 27 training training NOUN fcis-7187 78 28 sets set NOUN fcis-7187 78 29 are be AUX fcis-7187 78 30 increased increase VERB fcis-7187 78 31 [ [ X fcis-7187 78 32 8 8 NUM fcis-7187 78 33 ] ] PUNCT fcis-7187 78 34 . . PUNCT fcis-7187 79 1 figure figure NOUN fcis-7187 79 2 1 1 NUM fcis-7187 79 3 . . X fcis-7187 79 4 domain domain NOUN fcis-7187 79 5 - - PUNCT fcis-7187 79 6 adapted adapt VERB fcis-7187 79 7 classification classification NOUN fcis-7187 79 8 model model NOUN fcis-7187 79 9 of of ADP fcis-7187 79 10 tea tea NOUN fcis-7187 79 11 leaf leaf NOUN fcis-7187 79 12 disease disease NOUN fcis-7187 79 13 3.2 3.2 NUM fcis-7187 79 14 . . PUNCT fcis-7187 80 1 domain domain NOUN fcis-7187 80 2 - - PUNCT fcis-7187 80 3 adaptive adaptive NOUN fcis-7187 80 4 method method NOUN fcis-7187 80 5 the the DET fcis-7187 80 6 source source NOUN fcis-7187 80 7 domain domain NOUN fcis-7187 80 8 dataset dataset NOUN fcis-7187 80 9 in in ADP fcis-7187 80 10 this this DET fcis-7187 80 11 study study NOUN fcis-7187 80 12 contains contain VERB fcis-7187 80 13 a a DET fcis-7187 80 14 large large ADJ fcis-7187 80 15 amount amount NOUN fcis-7187 80 16 of of ADP fcis-7187 80 17 labeled label VERB fcis-7187 80 18 data datum NOUN fcis-7187 80 19 , , PUNCT fcis-7187 80 20 and and CCONJ fcis-7187 80 21 the the DET fcis-7187 80 22 target target NOUN fcis-7187 80 23 domain domain NOUN fcis-7187 80 24 dataset dataset NOUN fcis-7187 80 25 contains contain VERB fcis-7187 80 26 a a DET fcis-7187 80 27 large large ADJ fcis-7187 80 28 amount amount NOUN fcis-7187 80 29 of of ADP fcis-7187 80 30 label label NOUN fcis-7187 80 31 - - PUNCT fcis-7187 80 32 free free ADJ fcis-7187 80 33 data datum NOUN fcis-7187 80 34 . . PUNCT fcis-7187 81 1 assuming assume VERB fcis-7187 81 2 that that SCONJ fcis-7187 81 3 the the DET fcis-7187 81 4 data data NOUN fcis-7187 81 5 amount amount NOUN fcis-7187 81 6 of of ADP fcis-7187 81 7 the the DET fcis-7187 81 8 source source NOUN fcis-7187 81 9 domain domain NOUN fcis-7187 81 10 images image NOUN fcis-7187 81 11 x𝑠 x𝑠 INTJ fcis-7187 81 12 in in ADP fcis-7187 81 13 the the DET fcis-7187 81 14 tea tea NOUN fcis-7187 81 15 leaf leaf NOUN fcis-7187 81 16 disease disease NOUN fcis-7187 81 17 training training NOUN fcis-7187 81 18 set set NOUN fcis-7187 81 19 is be AUX fcis-7187 81 20 𝑛𝑠 𝑛𝑠 NOUN fcis-7187 81 21 , , PUNCT fcis-7187 81 22 and and CCONJ fcis-7187 81 23 the the DET fcis-7187 81 24 number number NOUN fcis-7187 81 25 of of ADP fcis-7187 81 26 categories category NOUN fcis-7187 81 27 is be AUX fcis-7187 81 28 k k PROPN fcis-7187 81 29 , , PUNCT fcis-7187 81 30 the the DET fcis-7187 81 31 source source NOUN fcis-7187 81 32 domain domain NOUN fcis-7187 81 33 data datum NOUN fcis-7187 81 34 can can AUX fcis-7187 81 35 be be AUX fcis-7187 81 36 expressed express VERB fcis-7187 81 37 as as ADP fcis-7187 81 38 𝐷𝑠 𝐷𝑠 PROPN fcis-7187 81 39 = = PUNCT fcis-7187 81 40 { { PUNCT fcis-7187 81 41 ( ( PUNCT fcis-7187 81 42 π‘₯𝑖 π‘₯𝑖 NUM fcis-7187 81 43 𝑠 𝑠 PROPN fcis-7187 81 44 , , PUNCT fcis-7187 81 45 𝑦𝑖 𝑦𝑖 PROPN fcis-7187 81 46 𝑠)}𝑖=1 𝑠)}𝑖=1 PROPN fcis-7187 81 47 𝑛𝑠 𝑛𝑠 VERB fcis-7187 81 48 , , PUNCT fcis-7187 81 49 where where SCONJ fcis-7187 81 50 π‘₯𝑠ϡx𝑠 π‘₯𝑠ϡx𝑠 NOUN fcis-7187 81 51 , , PUNCT fcis-7187 81 52 𝑦𝑠ϡy𝑠 𝑦𝑠ϡy𝑠 PROPN fcis-7187 81 53 = = SYM fcis-7187 81 54 { { PUNCT fcis-7187 81 55 1,2 1,2 NUM fcis-7187 81 56 , , PUNCT fcis-7187 81 57 … … PUNCT fcis-7187 81 58 , , PUNCT fcis-7187 81 59 𝐾 𝐾 NOUN fcis-7187 81 60 } } PUNCT fcis-7187 81 61 . . PUNCT fcis-7187 82 1 , , PUNCT fcis-7187 82 2 the the DET fcis-7187 82 3 amount amount NOUN fcis-7187 82 4 of of ADP fcis-7187 82 5 the the DET fcis-7187 82 6 target target NOUN fcis-7187 82 7 domain domain NOUN fcis-7187 82 8 images image NOUN fcis-7187 82 9 x𝑑 x𝑑 PRON fcis-7187 82 10 is be AUX fcis-7187 82 11 𝑛𝑑 𝑛𝑑 NUM fcis-7187 82 12 , , PUNCT fcis-7187 82 13 then then ADV fcis-7187 82 14 the the DET fcis-7187 82 15 target target NOUN fcis-7187 82 16 domain domain NOUN fcis-7187 82 17 data datum NOUN fcis-7187 82 18 can can AUX fcis-7187 82 19 be be AUX fcis-7187 82 20 represented represent VERB fcis-7187 82 21 as as ADP fcis-7187 82 22 𝐷𝑑 𝐷𝑑 PROPN fcis-7187 82 23 = = PUNCT fcis-7187 82 24 { { PUNCT fcis-7187 82 25 ( ( PUNCT fcis-7187 82 26 π‘₯𝑖 π‘₯𝑖 PROPN fcis-7187 82 27 𝑑)}𝑖=1 𝑑)}𝑖=1 NUM fcis-7187 82 28 𝑛𝑑 𝑛𝑑 INTJ fcis-7187 82 29 , , PUNCT fcis-7187 82 30 where where SCONJ fcis-7187 82 31 π‘₯𝑑ϡx𝑑 π‘₯𝑑ϡx𝑑 NOUN fcis-7187 82 32 . . PUNCT fcis-7187 83 1 the the DET fcis-7187 83 2 data data NOUN fcis-7187 83 3 distribution distribution NOUN fcis-7187 83 4 of of ADP fcis-7187 83 5 π‘₯𝑠 π‘₯𝑠 PROPN fcis-7187 83 6 and and CCONJ fcis-7187 83 7 π‘₯𝑑 π‘₯𝑑 ADV fcis-7187 83 8 are be AUX fcis-7187 83 9 similar similar ADJ fcis-7187 83 10 but but CCONJ fcis-7187 83 11 different different ADJ fcis-7187 83 12 , , PUNCT fcis-7187 83 13 constituting constitute VERB fcis-7187 83 14 the the DET fcis-7187 83 15 cross cross ADJ fcis-7187 83 16 - - ADJ fcis-7187 83 17 domain domain ADJ fcis-7187 83 18 dataset dataset NOUN fcis-7187 83 19 . . PUNCT fcis-7187 84 1 the the DET fcis-7187 84 2 data data NOUN fcis-7187 84 3 labels label VERB fcis-7187 84 4 π‘Œπ‘‘ π‘Œπ‘‘ PROPN fcis-7187 84 5 of of ADP fcis-7187 84 6 the the DET fcis-7187 84 7 dataset dataset NOUN fcis-7187 84 8 π‘₯𝑑 π‘₯𝑑 ADV fcis-7187 84 9 are be AUX fcis-7187 84 10 the the DET fcis-7187 84 11 same same ADJ fcis-7187 84 12 with with ADP fcis-7187 84 13 π‘Œπ‘  π‘Œπ‘  PROPN fcis-7187 84 14 . . PUNCT fcis-7187 85 1 the the DET fcis-7187 85 2 purpose purpose NOUN fcis-7187 85 3 of of ADP fcis-7187 85 4 domain domain NOUN fcis-7187 85 5 adaptation adaptation NOUN fcis-7187 85 6 is be AUX fcis-7187 85 7 to to PART fcis-7187 85 8 make make VERB fcis-7187 85 9 the the DET fcis-7187 85 10 network network NOUN fcis-7187 85 11 model model NOUN fcis-7187 85 12 with with ADP fcis-7187 85 13 the the DET fcis-7187 85 14 ability ability NOUN fcis-7187 85 15 of of ADP fcis-7187 85 16 source source NOUN fcis-7187 85 17 domain domain NOUN fcis-7187 85 18 data datum NOUN fcis-7187 85 19 classification classification NOUN fcis-7187 85 20 be be AUX fcis-7187 85 21 applied apply VERB fcis-7187 85 22 to to PART fcis-7187 85 23 predict predict VERB fcis-7187 85 24 the the DET fcis-7187 85 25 target target NOUN fcis-7187 85 26 domain domain NOUN fcis-7187 85 27 data datum NOUN fcis-7187 85 28 by by ADP fcis-7187 85 29 reducing reduce VERB fcis-7187 85 30 the the DET fcis-7187 85 31 difference difference NOUN fcis-7187 85 32 in in ADP fcis-7187 85 33 data data NOUN fcis-7187 85 34 distributions distribution NOUN fcis-7187 85 35 of of ADP fcis-7187 85 36 the the DET fcis-7187 85 37 source source NOUN fcis-7187 85 38 domain domain NOUN fcis-7187 85 39 and and CCONJ fcis-7187 85 40 target target NOUN fcis-7187 85 41 domain domain NOUN fcis-7187 85 42 . . PUNCT fcis-7187 86 1 the the DET fcis-7187 86 2 domain domain NOUN fcis-7187 86 3 - - PUNCT fcis-7187 86 4 adapted adapt VERB fcis-7187 86 5 classification classification NOUN fcis-7187 86 6 model model NOUN fcis-7187 86 7 of of ADP fcis-7187 86 8 tea tea NOUN fcis-7187 86 9 leaf leaf NOUN fcis-7187 86 10 disease disease NOUN fcis-7187 86 11 in in ADP fcis-7187 86 12 this this DET fcis-7187 86 13 study study NOUN fcis-7187 86 14 is be AUX fcis-7187 86 15 shown show VERB fcis-7187 86 16 in in ADP fcis-7187 86 17 figure figure NOUN fcis-7187 86 18 1 1 NUM fcis-7187 86 19 , , PUNCT fcis-7187 86 20 which which PRON fcis-7187 86 21 mainly mainly ADV fcis-7187 86 22 includes include VERB fcis-7187 86 23 three three NUM fcis-7187 86 24 modules module NOUN fcis-7187 86 25 : : PUNCT fcis-7187 86 26 feature feature NOUN fcis-7187 86 27 extraction extraction NOUN fcis-7187 86 28 module module NOUN fcis-7187 86 29 g g NOUN fcis-7187 86 30 , , PUNCT fcis-7187 86 31 leaf leaf NOUN fcis-7187 86 32 disease disease NOUN fcis-7187 86 33 classification classification NOUN fcis-7187 86 34 module module NOUN fcis-7187 86 35 c c NOUN fcis-7187 86 36 and and CCONJ fcis-7187 86 37 domain domain VERB fcis-7187 86 38 classification classification NOUN fcis-7187 86 39 module module NOUN fcis-7187 86 40 d. d. NOUN fcis-7187 86 41 the the DET fcis-7187 86 42 function function NOUN fcis-7187 86 43 of of ADP fcis-7187 86 44 the the DET fcis-7187 86 45 feature feature NOUN fcis-7187 86 46 extraction extraction NOUN fcis-7187 86 47 module module NOUN fcis-7187 86 48 is be AUX fcis-7187 86 49 to to PART fcis-7187 86 50 extract extract VERB fcis-7187 86 51 image image NOUN fcis-7187 86 52 features feature NOUN fcis-7187 86 53 and and CCONJ fcis-7187 86 54 confuse confuse VERB fcis-7187 86 55 the the DET fcis-7187 86 56 feature feature NOUN fcis-7187 86 57 distribution distribution NOUN fcis-7187 86 58 of of ADP fcis-7187 86 59 source source NOUN fcis-7187 86 60 domain domain NOUN fcis-7187 86 61 and and CCONJ fcis-7187 86 62 target target NOUN fcis-7187 86 63 domain domain NOUN fcis-7187 86 64 ; ; PUNCT fcis-7187 86 65 the the DET fcis-7187 86 66 leaf leaf NOUN fcis-7187 86 67 disease disease NOUN fcis-7187 86 68 classification classification NOUN fcis-7187 86 69 module module NOUN fcis-7187 86 70 is be AUX fcis-7187 86 71 to to PART fcis-7187 86 72 recognize recognize VERB fcis-7187 86 73 leaf leaf NOUN fcis-7187 86 74 disease disease NOUN fcis-7187 86 75 species specie NOUN fcis-7187 86 76 using use VERB fcis-7187 86 77 the the DET fcis-7187 86 78 extracted extract VERB fcis-7187 86 79 features feature NOUN fcis-7187 86 80 ; ; PUNCT fcis-7187 86 81 the the DET fcis-7187 86 82 domain domain NOUN fcis-7187 86 83 classification classification NOUN fcis-7187 86 84 module module NOUN fcis-7187 86 85 is be AUX fcis-7187 86 86 to to PART fcis-7187 86 87 recognize recognize VERB fcis-7187 86 88 which which DET fcis-7187 86 89 domains domain VERB fcis-7187 86 90 the the DET fcis-7187 86 91 input input NOUN fcis-7187 86 92 features feature NOUN fcis-7187 86 93 are be AUX fcis-7187 86 94 from from ADP fcis-7187 86 95 . . PUNCT fcis-7187 87 1 it it PRON fcis-7187 87 2 can can AUX fcis-7187 87 3 be be AUX fcis-7187 87 4 seen see VERB fcis-7187 87 5 that that SCONJ fcis-7187 87 6 the the DET fcis-7187 87 7 features feature NOUN fcis-7187 87 8 obtained obtain VERB fcis-7187 87 9 by by ADP fcis-7187 87 10 the the DET fcis-7187 87 11 feature feature NOUN fcis-7187 87 12 extraction extraction NOUN fcis-7187 87 13 module module NOUN fcis-7187 87 14 are be AUX fcis-7187 87 15 used use VERB fcis-7187 87 16 as as ADP fcis-7187 87 17 the the DET fcis-7187 87 18 input input NOUN fcis-7187 87 19 of of ADP fcis-7187 87 20 the the DET fcis-7187 87 21 other other ADJ fcis-7187 87 22 two two NUM fcis-7187 87 23 modules module NOUN fcis-7187 87 24 simultaneously simultaneously ADV fcis-7187 87 25 . . PUNCT fcis-7187 88 1 in in ADP fcis-7187 88 2 order order NOUN fcis-7187 88 3 to to PART fcis-7187 88 4 train train VERB fcis-7187 88 5 the the DET fcis-7187 88 6 model model NOUN fcis-7187 88 7 , , PUNCT fcis-7187 88 8 on on ADP fcis-7187 88 9 the the DET fcis-7187 88 10 one one NUM fcis-7187 88 11 hand hand NOUN fcis-7187 88 12 , , PUNCT fcis-7187 88 13 it it PRON fcis-7187 88 14 is be AUX fcis-7187 88 15 necessary necessary ADJ fcis-7187 88 16 to to PART fcis-7187 88 17 input input VERB fcis-7187 88 18 the the DET fcis-7187 88 19 source source NOUN fcis-7187 88 20 domain domain NOUN fcis-7187 88 21 images image NOUN fcis-7187 88 22 into into ADP fcis-7187 88 23 the the DET fcis-7187 88 24 feature feature NOUN fcis-7187 88 25 extraction extraction NOUN fcis-7187 88 26 module module NOUN fcis-7187 88 27 , , PUNCT fcis-7187 88 28 and and CCONJ fcis-7187 88 29 predict predict VERB fcis-7187 88 30 the the DET fcis-7187 88 31 leaf leaf NOUN fcis-7187 88 32 disease disease NOUN fcis-7187 88 33 type type NOUN fcis-7187 88 34 . . PUNCT fcis-7187 89 1 the the DET fcis-7187 89 2 classification classification NOUN fcis-7187 89 3 loss loss NOUN fcis-7187 89 4 𝐿𝑦 𝐿𝑦 PROPN fcis-7187 89 5 are be AUX fcis-7187 89 6 calculated calculate VERB fcis-7187 89 7 by by ADP fcis-7187 89 8 the the DET fcis-7187 89 9 predicted predict VERB fcis-7187 89 10 labels label NOUN fcis-7187 89 11 and and CCONJ fcis-7187 89 12 the the DET fcis-7187 89 13 true true ADJ fcis-7187 89 14 50 50 NUM fcis-7187 89 15 labels label NOUN fcis-7187 89 16 through through ADP fcis-7187 89 17 the the DET fcis-7187 89 18 cross cross ADJ fcis-7187 89 19 - - ADJ fcis-7187 89 20 entropy entropy ADJ fcis-7187 89 21 loss loss NOUN fcis-7187 89 22 function function NOUN fcis-7187 89 23 , , PUNCT fcis-7187 89 24 and and CCONJ fcis-7187 89 25 the the DET fcis-7187 89 26 parameters parameter NOUN fcis-7187 89 27 of of ADP fcis-7187 89 28 the the DET fcis-7187 89 29 network network NOUN fcis-7187 89 30 branch branch NOUN fcis-7187 89 31 are be AUX fcis-7187 89 32 updated update VERB fcis-7187 89 33 by by ADP fcis-7187 89 34 back back ADJ fcis-7187 89 35 propagation propagation NOUN fcis-7187 89 36 . . PUNCT fcis-7187 90 1 on on ADP fcis-7187 90 2 the the DET fcis-7187 90 3 other other ADJ fcis-7187 90 4 hand hand NOUN fcis-7187 90 5 , , PUNCT fcis-7187 90 6 it it PRON fcis-7187 90 7 is be AUX fcis-7187 90 8 necessary necessary ADJ fcis-7187 90 9 to to PART fcis-7187 90 10 input input VERB fcis-7187 90 11 the the DET fcis-7187 90 12 source source NOUN fcis-7187 90 13 domain domain NOUN fcis-7187 90 14 images image NOUN fcis-7187 90 15 and and CCONJ fcis-7187 90 16 the the DET fcis-7187 90 17 target target NOUN fcis-7187 90 18 domain domain NOUN fcis-7187 90 19 images image NOUN fcis-7187 90 20 at at ADP fcis-7187 90 21 the the DET fcis-7187 90 22 same same ADJ fcis-7187 90 23 time time NOUN fcis-7187 90 24 . . PUNCT fcis-7187 91 1 then then ADV fcis-7187 91 2 the the DET fcis-7187 91 3 features feature NOUN fcis-7187 91 4 of of ADP fcis-7187 91 5 these these DET fcis-7187 91 6 images image NOUN fcis-7187 91 7 are be AUX fcis-7187 91 8 extracted extract VERB fcis-7187 91 9 to to PART fcis-7187 91 10 enter enter VERB fcis-7187 91 11 the the DET fcis-7187 91 12 domain domain NOUN fcis-7187 91 13 classification classification NOUN fcis-7187 91 14 module module NOUN fcis-7187 91 15 for for ADP fcis-7187 91 16 domain domain NOUN fcis-7187 91 17 prediction prediction NOUN fcis-7187 91 18 . . PUNCT fcis-7187 92 1 the the DET fcis-7187 92 2 domain domain NOUN fcis-7187 92 3 prediction prediction NOUN fcis-7187 92 4 loss loss NOUN fcis-7187 92 5 of of ADP fcis-7187 92 6 this this DET fcis-7187 92 7 branch branch NOUN fcis-7187 92 8 can can AUX fcis-7187 92 9 be be AUX fcis-7187 92 10 expressed express VERB fcis-7187 92 11 as as ADP fcis-7187 92 12 : : PUNCT fcis-7187 92 13 𝐿𝑑 𝐿𝑑 PROPN fcis-7187 92 14 = = PROPN fcis-7187 92 15 βˆ’βˆ‘π‘‘log βˆ’βˆ‘π‘‘log PROPN fcis-7187 92 16 ( ( PUNCT fcis-7187 92 17 𝐷(𝐺(𝐼 𝐷(𝐺(𝐼 PROPN fcis-7187 92 18 ) ) PUNCT fcis-7187 92 19 ) ) PUNCT fcis-7187 92 20 ) ) PUNCT fcis-7187 93 1 𝐼 𝐼 ADP fcis-7187 93 2 βˆ’βˆ‘(1 βˆ’βˆ‘(1 NUM fcis-7187 93 3 βˆ’ βˆ’ PROPN fcis-7187 93 4 𝑑)log(1 𝑑)log(1 PUNCT fcis-7187 93 5 βˆ’ βˆ’ NOUN fcis-7187 93 6 𝐷(𝐺(𝐼 𝐷(𝐺(𝐼 NOUN fcis-7187 93 7 ) ) PUNCT fcis-7187 93 8 ) ) PUNCT fcis-7187 93 9 ) ) PUNCT fcis-7187 94 1 𝐼 𝐼 ADP fcis-7187 94 2 where where SCONJ fcis-7187 94 3 𝐼 𝐼 PROPN fcis-7187 94 4 represents represent VERB fcis-7187 94 5 the the DET fcis-7187 94 6 images image NOUN fcis-7187 94 7 from from ADP fcis-7187 94 8 source source NOUN fcis-7187 94 9 and and CCONJ fcis-7187 94 10 target target NOUN fcis-7187 94 11 domain domain NOUN fcis-7187 94 12 , , PUNCT fcis-7187 94 13 d d PROPN fcis-7187 94 14 takes take VERB fcis-7187 94 15 the the DET fcis-7187 94 16 value value NOUN fcis-7187 94 17 0 0 PUNCT fcis-7187 94 18 when when SCONJ fcis-7187 94 19 the the DET fcis-7187 94 20 input input NOUN fcis-7187 94 21 image image NOUN fcis-7187 94 22 is be AUX fcis-7187 94 23 from from ADP fcis-7187 94 24 the the DET fcis-7187 94 25 source source NOUN fcis-7187 94 26 domain domain NOUN fcis-7187 94 27 and and CCONJ fcis-7187 94 28 1 1 NUM fcis-7187 94 29 when when SCONJ fcis-7187 94 30 it it PRON fcis-7187 94 31 is be AUX fcis-7187 94 32 from from ADP fcis-7187 94 33 the the DET fcis-7187 94 34 target target NOUN fcis-7187 94 35 domain domain NOUN fcis-7187 94 36 . . PUNCT fcis-7187 95 1 the the DET fcis-7187 95 2 training training NOUN fcis-7187 95 3 goal goal NOUN fcis-7187 95 4 of of ADP fcis-7187 95 5 the the DET fcis-7187 95 6 domain domain NOUN fcis-7187 95 7 classification classification NOUN fcis-7187 95 8 module module NOUN fcis-7187 95 9 is be AUX fcis-7187 95 10 to to PART fcis-7187 95 11 make make VERB fcis-7187 95 12 the the DET fcis-7187 95 13 domain domain NOUN fcis-7187 95 14 classifier classifier NOUN fcis-7187 95 15 more more ADV fcis-7187 95 16 accurate accurate ADJ fcis-7187 95 17 , , PUNCT fcis-7187 95 18 while while SCONJ fcis-7187 95 19 the the DET fcis-7187 95 20 training training NOUN fcis-7187 95 21 goal goal NOUN fcis-7187 95 22 of of ADP fcis-7187 95 23 the the DET fcis-7187 95 24 feature feature NOUN fcis-7187 95 25 extraction extraction NOUN fcis-7187 95 26 module module NOUN fcis-7187 95 27 is be AUX fcis-7187 95 28 to to PART fcis-7187 95 29 extract extract VERB fcis-7187 95 30 the the DET fcis-7187 95 31 common common ADJ fcis-7187 95 32 features feature NOUN fcis-7187 95 33 in in ADP fcis-7187 95 34 the the DET fcis-7187 95 35 source source NOUN fcis-7187 95 36 domain domain NOUN fcis-7187 95 37 and and CCONJ fcis-7187 95 38 the the DET fcis-7187 95 39 target target NOUN fcis-7187 95 40 domain domain NOUN fcis-7187 95 41 , , PUNCT fcis-7187 95 42 making make VERB fcis-7187 95 43 the the DET fcis-7187 95 44 domain domain NOUN fcis-7187 95 45 classifier classifier NOUN fcis-7187 95 46 unable unable ADJ fcis-7187 95 47 to to PART fcis-7187 95 48 discriminate discriminate VERB fcis-7187 95 49 . . PUNCT fcis-7187 96 1 therefore therefore ADV fcis-7187 96 2 , , PUNCT fcis-7187 96 3 the the DET fcis-7187 96 4 network network NOUN fcis-7187 96 5 training training NOUN fcis-7187 96 6 in in ADP fcis-7187 96 7 this this DET fcis-7187 96 8 branch branch NOUN fcis-7187 96 9 is be AUX fcis-7187 96 10 a a DET fcis-7187 96 11 typical typical ADJ fcis-7187 96 12 adversarial adversarial ADJ fcis-7187 96 13 thinking thinking NOUN fcis-7187 96 14 , , PUNCT fcis-7187 96 15 and and CCONJ fcis-7187 96 16 its its PRON fcis-7187 96 17 training training NOUN fcis-7187 96 18 goal goal NOUN fcis-7187 96 19 can can AUX fcis-7187 96 20 be be AUX fcis-7187 96 21 expressed express VERB fcis-7187 96 22 as as ADP fcis-7187 96 23 max max PROPN fcis-7187 96 24 𝐺 𝐺 PROPN fcis-7187 96 25 min min PROPN fcis-7187 96 26 𝐷 𝐷 PROPN fcis-7187 96 27 𝐿𝑑. 𝐿𝑑. PROPN fcis-7187 96 28 in in ADP fcis-7187 96 29 the the DET fcis-7187 96 30 specific specific ADJ fcis-7187 96 31 implementation implementation NOUN fcis-7187 96 32 , , PUNCT fcis-7187 96 33 gradient gradient ADJ fcis-7187 96 34 reversal reversal NOUN fcis-7187 96 35 layer layer NOUN fcis-7187 96 36 ( ( PUNCT fcis-7187 96 37 grl grl PROPN fcis-7187 96 38 ) ) PUNCT fcis-7187 97 1 [ [ X fcis-7187 97 2 9 9 NUM fcis-7187 97 3 ] ] PUNCT fcis-7187 97 4 is be AUX fcis-7187 97 5 added add VERB fcis-7187 97 6 between between ADP fcis-7187 97 7 the the DET fcis-7187 97 8 feature feature NOUN fcis-7187 97 9 extraction extraction NOUN fcis-7187 97 10 module module NOUN fcis-7187 97 11 and and CCONJ fcis-7187 97 12 the the DET fcis-7187 97 13 domain domain NOUN fcis-7187 97 14 classification classification NOUN fcis-7187 97 15 module module NOUN fcis-7187 97 16 to to PART fcis-7187 97 17 achieve achieve VERB fcis-7187 97 18 the the DET fcis-7187 97 19 effect effect NOUN fcis-7187 97 20 of of ADP fcis-7187 97 21 adversarial adversarial ADJ fcis-7187 97 22 training training NOUN fcis-7187 97 23 . . PUNCT fcis-7187 98 1 the the DET fcis-7187 98 2 loss loss NOUN fcis-7187 98 3 function function NOUN fcis-7187 98 4 of of ADP fcis-7187 98 5 the the DET fcis-7187 98 6 above above ADJ fcis-7187 98 7 two two NUM fcis-7187 98 8 branches branch NOUN fcis-7187 98 9 constitutes constitute VERB fcis-7187 98 10 the the DET fcis-7187 98 11 final final ADJ fcis-7187 98 12 loss loss NOUN fcis-7187 98 13 of of ADP fcis-7187 98 14 the the DET fcis-7187 98 15 network network NOUN fcis-7187 98 16 model model NOUN fcis-7187 98 17 , , PUNCT fcis-7187 98 18 which which PRON fcis-7187 98 19 is be AUX fcis-7187 98 20 expressed express VERB fcis-7187 98 21 as as ADP fcis-7187 98 22 πΏπ‘‘π‘œπ‘‘π‘Žπ‘™ πΏπ‘‘π‘œπ‘‘π‘Žπ‘™ PROPN fcis-7187 98 23 = = PUNCT fcis-7187 98 24 𝐿𝑦 𝐿𝑦 PROPN fcis-7187 98 25 + + CCONJ fcis-7187 98 26 πœ†πΏπ‘‘πœ† πœ†πΏπ‘‘πœ† PROPN fcis-7187 98 27 , , PUNCT fcis-7187 98 28 where where SCONJ fcis-7187 98 29 πœ† πœ† NOUN fcis-7187 98 30 is be AUX fcis-7187 98 31 the the DET fcis-7187 98 32 hyperparameter hyperparameter NOUN fcis-7187 98 33 used use VERB fcis-7187 98 34 to to PART fcis-7187 98 35 balance balance VERB fcis-7187 98 36 the the DET fcis-7187 98 37 two two NUM fcis-7187 98 38 branches branch NOUN fcis-7187 98 39 of of ADP fcis-7187 98 40 the the DET fcis-7187 98 41 network network NOUN fcis-7187 98 42 . . PUNCT fcis-7187 99 1 after after ADP fcis-7187 99 2 the the DET fcis-7187 99 3 training training NOUN fcis-7187 99 4 stage stage NOUN fcis-7187 99 5 , , PUNCT fcis-7187 99 6 tea tea NOUN fcis-7187 99 7 leaf leaf NOUN fcis-7187 99 8 disease disease NOUN fcis-7187 99 9 types type NOUN fcis-7187 99 10 are be AUX fcis-7187 99 11 predicted predict VERB fcis-7187 99 12 using use VERB fcis-7187 99 13 the the DET fcis-7187 99 14 test test NOUN fcis-7187 99 15 set set VERB fcis-7187 99 16 data datum NOUN fcis-7187 99 17 . . PUNCT fcis-7187 100 1 at at ADP fcis-7187 100 2 this this DET fcis-7187 100 3 time time NOUN fcis-7187 100 4 , , PUNCT fcis-7187 100 5 the the DET fcis-7187 100 6 domain domain NOUN fcis-7187 100 7 classification classification NOUN fcis-7187 100 8 module module NOUN fcis-7187 100 9 is be AUX fcis-7187 100 10 removed remove VERB fcis-7187 100 11 , , PUNCT fcis-7187 100 12 and and CCONJ fcis-7187 100 13 only only ADV fcis-7187 100 14 the the DET fcis-7187 100 15 feature feature NOUN fcis-7187 100 16 extraction extraction NOUN fcis-7187 100 17 module module NOUN fcis-7187 100 18 and and CCONJ fcis-7187 100 19 leaf leaf NOUN fcis-7187 100 20 disease disease NOUN fcis-7187 100 21 classification classification NOUN fcis-7187 100 22 module module NOUN fcis-7187 100 23 are be AUX fcis-7187 100 24 reserved reserve VERB fcis-7187 100 25 for for ADP fcis-7187 100 26 prediction prediction NOUN fcis-7187 100 27 . . PUNCT fcis-7187 101 1 table table NOUN fcis-7187 101 2 1 1 NUM fcis-7187 101 3 . . PUNCT fcis-7187 101 4 comparison comparison NOUN fcis-7187 101 5 of of ADP fcis-7187 101 6 classification classification NOUN fcis-7187 101 7 accuracy accuracy NOUN fcis-7187 101 8 of of ADP fcis-7187 101 9 tea tea NOUN fcis-7187 101 10 leaf leaf NOUN fcis-7187 101 11 disease disease NOUN fcis-7187 101 12 based base VERB fcis-7187 101 13 on on ADP fcis-7187 101 14 different different ADJ fcis-7187 101 15 models model NOUN fcis-7187 101 16 method method NOUN fcis-7187 101 17 accuracy accuracy NOUN fcis-7187 101 18 traditional traditional ADJ fcis-7187 101 19 neural neural ADJ fcis-7187 101 20 network network NOUN fcis-7187 101 21 model model NOUN fcis-7187 101 22 46.1 46.1 NUM fcis-7187 101 23 % % NOUN fcis-7187 101 24 domain domain NOUN fcis-7187 101 25 adaptive adaptive ADJ fcis-7187 101 26 model model NOUN fcis-7187 101 27 84.7 84.7 NUM fcis-7187 101 28 % % NOUN fcis-7187 101 29 4 4 NUM fcis-7187 101 30 . . NOUN fcis-7187 101 31 experiment experiment NOUN fcis-7187 101 32 to to PART fcis-7187 101 33 verify verify VERB fcis-7187 101 34 the the DET fcis-7187 101 35 effectiveness effectiveness NOUN fcis-7187 101 36 of of ADP fcis-7187 101 37 the the DET fcis-7187 101 38 domain domain NOUN fcis-7187 101 39 adaptive adaptive ADJ fcis-7187 101 40 method method NOUN fcis-7187 101 41 in in ADP fcis-7187 101 42 this this DET fcis-7187 101 43 study study NOUN fcis-7187 101 44 , , PUNCT fcis-7187 101 45 we we PRON fcis-7187 101 46 compare compare VERB fcis-7187 101 47 it it PRON fcis-7187 101 48 with with ADP fcis-7187 101 49 the the DET fcis-7187 101 50 traditional traditional ADJ fcis-7187 101 51 neural neural ADJ fcis-7187 101 52 network network NOUN fcis-7187 101 53 model model NOUN fcis-7187 101 54 method method NOUN fcis-7187 101 55 . . PUNCT fcis-7187 102 1 both both CCONJ fcis-7187 102 2 the the DET fcis-7187 102 3 feature feature NOUN fcis-7187 102 4 extraction extraction NOUN fcis-7187 102 5 module module NOUN fcis-7187 102 6 and and CCONJ fcis-7187 102 7 the the DET fcis-7187 102 8 leaf leaf NOUN fcis-7187 102 9 disease disease NOUN fcis-7187 102 10 classification classification NOUN fcis-7187 102 11 module module NOUN fcis-7187 102 12 use use VERB fcis-7187 102 13 a a DET fcis-7187 102 14 pre pre ADJ fcis-7187 102 15 - - ADJ fcis-7187 102 16 trained train VERB fcis-7187 102 17 googlenet googlenet NOUN fcis-7187 102 18 neural neural ADJ fcis-7187 102 19 network network NOUN fcis-7187 102 20 architecture architecture NOUN fcis-7187 102 21 . . PUNCT fcis-7187 103 1 the the DET fcis-7187 103 2 domain domain NOUN fcis-7187 103 3 classification classification NOUN fcis-7187 103 4 module module NOUN fcis-7187 103 5 consists consist VERB fcis-7187 103 6 of of ADP fcis-7187 103 7 three three NUM fcis-7187 103 8 fully fully ADV fcis-7187 103 9 connected connected ADJ fcis-7187 103 10 layers layer NOUN fcis-7187 103 11 with with ADP fcis-7187 103 12 a a DET fcis-7187 103 13 number number NOUN fcis-7187 103 14 of of ADP fcis-7187 103 15 neurons neuron NOUN fcis-7187 103 16 of of ADP fcis-7187 103 17 1024 1024 NUM fcis-7187 103 18 , , PUNCT fcis-7187 103 19 512 512 NUM fcis-7187 103 20 and and CCONJ fcis-7187 103 21 1 1 NUM fcis-7187 103 22 , , PUNCT fcis-7187 103 23 respectively respectively ADV fcis-7187 103 24 . . PUNCT fcis-7187 104 1 the the DET fcis-7187 104 2 experimental experimental ADJ fcis-7187 104 3 comparison comparison NOUN fcis-7187 104 4 results result NOUN fcis-7187 104 5 are be AUX fcis-7187 104 6 shown show VERB fcis-7187 104 7 in in ADP fcis-7187 104 8 table table NOUN fcis-7187 104 9 1 1 NUM fcis-7187 104 10 . . PUNCT fcis-7187 105 1 it it PRON fcis-7187 105 2 can can AUX fcis-7187 105 3 be be AUX fcis-7187 105 4 seen see VERB fcis-7187 105 5 that that SCONJ fcis-7187 105 6 the the DET fcis-7187 105 7 average average ADJ fcis-7187 105 8 accuracy accuracy NOUN fcis-7187 105 9 of of ADP fcis-7187 105 10 cross cross ADJ fcis-7187 105 11 - - ADJ fcis-7187 105 12 domain domain ADJ fcis-7187 105 13 classification classification NOUN fcis-7187 105 14 of of ADP fcis-7187 105 15 tea tea NOUN fcis-7187 105 16 leaf leaf NOUN fcis-7187 105 17 disease disease NOUN fcis-7187 105 18 in in ADP fcis-7187 105 19 the the DET fcis-7187 105 20 traditional traditional ADJ fcis-7187 105 21 neural neural ADJ fcis-7187 105 22 network network NOUN fcis-7187 105 23 model model NOUN fcis-7187 105 24 trained train VERB fcis-7187 105 25 only only ADV fcis-7187 105 26 with with ADP fcis-7187 105 27 source source NOUN fcis-7187 105 28 domain domain NOUN fcis-7187 105 29 data datum NOUN fcis-7187 105 30 is be AUX fcis-7187 105 31 only only ADV fcis-7187 105 32 46.1 46.1 NUM fcis-7187 105 33 % % NOUN fcis-7187 105 34 , , PUNCT fcis-7187 105 35 indicating indicate VERB fcis-7187 105 36 that that SCONJ fcis-7187 105 37 there there PRON fcis-7187 105 38 is be VERB fcis-7187 105 39 a a DET fcis-7187 105 40 large large ADJ fcis-7187 105 41 data datum NOUN fcis-7187 105 42 distribution distribution NOUN fcis-7187 105 43 difference difference NOUN fcis-7187 105 44 between between ADP fcis-7187 105 45 the the DET fcis-7187 105 46 target target NOUN fcis-7187 105 47 domain domain NOUN fcis-7187 105 48 image image NOUN fcis-7187 105 49 and and CCONJ fcis-7187 105 50 the the DET fcis-7187 105 51 source source NOUN fcis-7187 105 52 domain domain NOUN fcis-7187 105 53 image image NOUN fcis-7187 105 54 after after ADP fcis-7187 105 55 brightness brightness NOUN fcis-7187 105 56 processing processing NOUN fcis-7187 105 57 , , PUNCT fcis-7187 105 58 and and CCONJ fcis-7187 105 59 the the DET fcis-7187 105 60 classification classification NOUN fcis-7187 105 61 accuracy accuracy NOUN fcis-7187 105 62 is be AUX fcis-7187 105 63 low low ADJ fcis-7187 105 64 . . PUNCT fcis-7187 106 1 however however ADV fcis-7187 106 2 , , PUNCT fcis-7187 106 3 the the DET fcis-7187 106 4 neural neural ADJ fcis-7187 106 5 network network NOUN fcis-7187 106 6 trained train VERB fcis-7187 106 7 by by ADP fcis-7187 106 8 domain domain NOUN fcis-7187 106 9 adaptation adaptation NOUN fcis-7187 106 10 can can AUX fcis-7187 106 11 reduce reduce VERB fcis-7187 106 12 the the DET fcis-7187 106 13 difference difference NOUN fcis-7187 106 14 in in ADP fcis-7187 106 15 data datum NOUN fcis-7187 106 16 distribution distribution NOUN fcis-7187 106 17 , , PUNCT fcis-7187 106 18 and and CCONJ fcis-7187 106 19 improve improve VERB fcis-7187 106 20 the the DET fcis-7187 106 21 classification classification NOUN fcis-7187 106 22 accuracy accuracy NOUN fcis-7187 106 23 of of ADP fcis-7187 106 24 the the DET fcis-7187 106 25 test test NOUN fcis-7187 106 26 set set VERB fcis-7187 106 27 to to ADP fcis-7187 106 28 84.7 84.7 NUM fcis-7187 106 29 % % NOUN fcis-7187 106 30 . . PUNCT fcis-7187 107 1 the the DET fcis-7187 107 2 experimental experimental ADJ fcis-7187 107 3 results result NOUN fcis-7187 107 4 fully fully ADV fcis-7187 107 5 illustrate illustrate VERB fcis-7187 107 6 the the DET fcis-7187 107 7 necessity necessity NOUN fcis-7187 107 8 of of ADP fcis-7187 107 9 using use VERB fcis-7187 107 10 the the DET fcis-7187 107 11 domain domain NOUN fcis-7187 107 12 adaptive adaptive ADJ fcis-7187 107 13 algorithm algorithm NOUN fcis-7187 107 14 on on ADP fcis-7187 107 15 the the DET fcis-7187 107 16 cross cross ADJ fcis-7187 107 17 - - ADJ fcis-7187 107 18 domain domain ADJ fcis-7187 107 19 data datum NOUN fcis-7187 107 20 set set VERB fcis-7187 107 21 of of ADP fcis-7187 107 22 tea tea NOUN fcis-7187 107 23 leaf leaf NOUN fcis-7187 107 24 disease disease NOUN fcis-7187 107 25 , , PUNCT fcis-7187 107 26 which which PRON fcis-7187 107 27 can can AUX fcis-7187 107 28 bring bring VERB fcis-7187 107 29 great great ADJ fcis-7187 107 30 accuracy accuracy NOUN fcis-7187 107 31 improvement improvement NOUN fcis-7187 107 32 . . PUNCT fcis-7187 108 1 5 5 X fcis-7187 108 2 . . X fcis-7187 108 3 conclusion conclusion NOUN fcis-7187 108 4 in in ADP fcis-7187 108 5 this this DET fcis-7187 108 6 paper paper NOUN fcis-7187 108 7 , , PUNCT fcis-7187 108 8 the the DET fcis-7187 108 9 classification classification NOUN fcis-7187 108 10 method method NOUN fcis-7187 108 11 of of ADP fcis-7187 108 12 tea tea NOUN fcis-7187 108 13 leaf leaf NOUN fcis-7187 108 14 disease disease NOUN fcis-7187 108 15 based base VERB fcis-7187 108 16 on on ADP fcis-7187 108 17 domain domain NOUN fcis-7187 108 18 adaptation adaptation NOUN fcis-7187 108 19 are be AUX fcis-7187 108 20 proposed propose VERB fcis-7187 108 21 . . PUNCT fcis-7187 109 1 at at ADP fcis-7187 109 2 first first ADV fcis-7187 109 3 the the DET fcis-7187 109 4 tea tea NOUN fcis-7187 109 5 leaf leaf NOUN fcis-7187 109 6 disease disease NOUN fcis-7187 109 7 classifier classifier NOUN fcis-7187 109 8 is be AUX fcis-7187 109 9 trained train VERB fcis-7187 109 10 with with ADP fcis-7187 109 11 source source NOUN fcis-7187 109 12 domain domain NOUN fcis-7187 109 13 data datum NOUN fcis-7187 109 14 . . PUNCT fcis-7187 110 1 then then ADV fcis-7187 110 2 the the DET fcis-7187 110 3 data data NOUN fcis-7187 110 4 distribution distribution NOUN fcis-7187 110 5 between between ADP fcis-7187 110 6 different different ADJ fcis-7187 110 7 domains domain NOUN fcis-7187 110 8 are be AUX fcis-7187 110 9 aligned align VERB fcis-7187 110 10 through through ADP fcis-7187 110 11 adversarial adversarial ADJ fcis-7187 110 12 training training NOUN fcis-7187 110 13 , , PUNCT fcis-7187 110 14 so so SCONJ fcis-7187 110 15 as as SCONJ fcis-7187 110 16 to to PART fcis-7187 110 17 achieve achieve VERB fcis-7187 110 18 a a DET fcis-7187 110 19 good good ADJ fcis-7187 110 20 classification classification NOUN fcis-7187 110 21 effect effect NOUN fcis-7187 110 22 of of ADP fcis-7187 110 23 leaf leaf NOUN fcis-7187 110 24 disease disease NOUN fcis-7187 110 25 on on ADP fcis-7187 110 26 cross cross ADJ fcis-7187 110 27 - - ADJ fcis-7187 110 28 domain domain ADJ fcis-7187 110 29 datasets dataset NOUN fcis-7187 110 30 . . PUNCT fcis-7187 111 1 the the DET fcis-7187 111 2 experimental experimental ADJ fcis-7187 111 3 results result NOUN fcis-7187 111 4 verify verify VERB fcis-7187 111 5 the the DET fcis-7187 111 6 effectiveness effectiveness NOUN fcis-7187 111 7 of of ADP fcis-7187 111 8 the the DET fcis-7187 111 9 domain domain NOUN fcis-7187 111 10 adaptive adaptive ADJ fcis-7187 111 11 method method NOUN fcis-7187 111 12 to to PART fcis-7187 111 13 solve solve VERB fcis-7187 111 14 the the DET fcis-7187 111 15 cross cross ADJ fcis-7187 111 16 - - ADJ fcis-7187 111 17 domain domain ADJ fcis-7187 111 18 classification classification NOUN fcis-7187 111 19 problem problem NOUN fcis-7187 111 20 of of ADP fcis-7187 111 21 tea tea NOUN fcis-7187 111 22 leaf leaf NOUN fcis-7187 111 23 disease disease NOUN fcis-7187 111 24 . . PUNCT fcis-7187 112 1 future future ADJ fcis-7187 112 2 studies study NOUN fcis-7187 112 3 will will AUX fcis-7187 112 4 combine combine VERB fcis-7187 112 5 cross cross ADJ fcis-7187 112 6 - - ADJ fcis-7187 112 7 domain domain ADJ fcis-7187 112 8 classification classification NOUN fcis-7187 112 9 and and CCONJ fcis-7187 112 10 few few ADJ fcis-7187 112 11 - - PUNCT fcis-7187 112 12 shot shot NOUN fcis-7187 112 13 learning learning NOUN fcis-7187 112 14 to to PART fcis-7187 112 15 conduct conduct VERB fcis-7187 112 16 more more ADV fcis-7187 112 17 beneficial beneficial ADJ fcis-7187 112 18 exploration exploration NOUN fcis-7187 112 19 in in ADP fcis-7187 112 20 the the DET fcis-7187 112 21 field field NOUN fcis-7187 112 22 of of ADP fcis-7187 112 23 smart smart ADJ fcis-7187 112 24 agriculture agriculture NOUN fcis-7187 112 25 . . PUNCT fcis-7187 113 1 acknowledgment acknowledgment NOUN fcis-7187 113 2 this this DET fcis-7187 113 3 work work NOUN fcis-7187 113 4 was be AUX fcis-7187 113 5 supported support VERB fcis-7187 113 6 by by ADP fcis-7187 113 7 the the DET fcis-7187 113 8 tai'an tai'an PROPN fcis-7187 113 9 science science NOUN fcis-7187 113 10 and and CCONJ fcis-7187 113 11 technology technology NOUN fcis-7187 113 12 innovation innovation NOUN fcis-7187 113 13 development development NOUN fcis-7187 113 14 project project NOUN fcis-7187 113 15 under under ADP fcis-7187 113 16 grant grant NOUN fcis-7187 113 17 no.2021ns097 no.2021ns097 PROPN fcis-7187 113 18 . . PUNCT fcis-7187 114 1 references reference NOUN fcis-7187 114 2 [ [ X fcis-7187 114 3 1 1 NUM fcis-7187 114 4 ] ] X fcis-7187 114 5 chen chen PROPN fcis-7187 114 6 j j PROPN fcis-7187 114 7 , , PUNCT fcis-7187 114 8 liu liu PROPN fcis-7187 114 9 q q PROPN fcis-7187 114 10 , , PUNCT fcis-7187 114 11 gao gao PROPN fcis-7187 114 12 l. l. PROPN fcis-7187 114 13 visual visual PROPN fcis-7187 114 14 tea tea NOUN fcis-7187 114 15 leaf leaf NOUN fcis-7187 114 16 disease disease NOUN fcis-7187 114 17 recognition recognition NOUN fcis-7187 114 18 using use VERB fcis-7187 114 19 a a DET fcis-7187 114 20 convolutional convolutional ADJ fcis-7187 114 21 neural neural ADJ fcis-7187 114 22 network network NOUN fcis-7187 114 23 model[j model[j PROPN fcis-7187 114 24 ] ] PUNCT fcis-7187 114 25 . . PUNCT fcis-7187 115 1 symmetry symmetry PROPN fcis-7187 115 2 , , PUNCT fcis-7187 115 3 2019 2019 NUM fcis-7187 115 4 , , PUNCT fcis-7187 115 5 11(3 11(3 NUM fcis-7187 115 6 ): ): PUNCT fcis-7187 115 7 343 343 NUM fcis-7187 115 8 . . PUNCT fcis-7187 116 1 [ [ X fcis-7187 116 2 2 2 X fcis-7187 116 3 ] ] X fcis-7187 116 4 gayathri gayathri PROPN fcis-7187 116 5 s s PROPN fcis-7187 116 6 , , PUNCT fcis-7187 116 7 wise wise ADJ fcis-7187 116 8 d d X fcis-7187 116 9 c c PROPN fcis-7187 116 10 j j PROPN fcis-7187 116 11 w w PROPN fcis-7187 116 12 , , PUNCT fcis-7187 116 13 shamini shamini PROPN fcis-7187 116 14 p p PROPN fcis-7187 116 15 b b PROPN fcis-7187 116 16 , , PUNCT fcis-7187 116 17 et et PROPN fcis-7187 116 18 al al PROPN fcis-7187 116 19 . . PUNCT fcis-7187 117 1 image image NOUN fcis-7187 117 2 analysis analysis NOUN fcis-7187 117 3 and and CCONJ fcis-7187 117 4 detection detection NOUN fcis-7187 117 5 of of ADP fcis-7187 117 6 tea tea NOUN fcis-7187 117 7 leaf leaf NOUN fcis-7187 117 8 disease disease NOUN fcis-7187 117 9 using use VERB fcis-7187 117 10 deep deep ADJ fcis-7187 117 11 learning[c]//2020 learning[c]//2020 PROPN fcis-7187 117 12 international international ADJ fcis-7187 117 13 conference conference NOUN fcis-7187 117 14 on on ADP fcis-7187 117 15 electronics electronic NOUN fcis-7187 117 16 and and CCONJ fcis-7187 117 17 sustainable sustainable ADJ fcis-7187 117 18 communication communication NOUN fcis-7187 117 19 systems system NOUN fcis-7187 117 20 ( ( PUNCT fcis-7187 117 21 icesc icesc PROPN fcis-7187 117 22 ) ) PUNCT fcis-7187 117 23 . . PUNCT fcis-7187 118 1 ieee ieee PROPN fcis-7187 118 2 , , PUNCT fcis-7187 118 3 2020 2020 NUM fcis-7187 118 4 : : PUNCT fcis-7187 118 5 398 398 NUM fcis-7187 118 6 - - SYM fcis-7187 118 7 403 403 NUM fcis-7187 118 8 . . PUNCT fcis-7187 119 1 [ [ X fcis-7187 119 2 3 3 NUM fcis-7187 119 3 ] ] SYM fcis-7187 119 4 hu hu PROPN fcis-7187 119 5 g g PROPN fcis-7187 119 6 , , PUNCT fcis-7187 119 7 yang yang PROPN fcis-7187 119 8 x x PROPN fcis-7187 119 9 , , PUNCT fcis-7187 119 10 zhang zhang PROPN fcis-7187 119 11 y y PROPN fcis-7187 119 12 , , PUNCT fcis-7187 119 13 et et PROPN fcis-7187 119 14 al al PROPN fcis-7187 119 15 . . PUNCT fcis-7187 119 16 identification identification NOUN fcis-7187 119 17 of of ADP fcis-7187 119 18 tea tea NOUN fcis-7187 119 19 leaf leaf NOUN fcis-7187 119 20 diseases disease NOUN fcis-7187 119 21 by by ADP fcis-7187 119 22 using use VERB fcis-7187 119 23 an an DET fcis-7187 119 24 improved improved ADJ fcis-7187 119 25 deep deep ADJ fcis-7187 119 26 convolutional convolutional ADJ fcis-7187 119 27 neural neural ADJ fcis-7187 119 28 network[j network[j NOUN fcis-7187 119 29 ] ] PUNCT fcis-7187 119 30 . . PUNCT fcis-7187 120 1 sustainable sustainable ADJ fcis-7187 120 2 computing computing NOUN fcis-7187 120 3 : : PUNCT fcis-7187 120 4 informatics informatic NOUN fcis-7187 120 5 and and CCONJ fcis-7187 120 6 systems system NOUN fcis-7187 120 7 , , PUNCT fcis-7187 120 8 2019 2019 NUM fcis-7187 120 9 , , PUNCT fcis-7187 120 10 24 24 NUM fcis-7187 120 11 : : SYM fcis-7187 120 12 100353 100353 NUM fcis-7187 120 13 . . PUNCT fcis-7187 121 1 [ [ X fcis-7187 121 2 4 4 NUM fcis-7187 121 3 ] ] X fcis-7187 121 4 ganin ganin PROPN fcis-7187 121 5 y y PROPN fcis-7187 121 6 , , PUNCT fcis-7187 121 7 ustinova ustinova X fcis-7187 121 8 e e NOUN fcis-7187 121 9 , , PUNCT fcis-7187 121 10 ajakan ajakan ADJ fcis-7187 121 11 h h NOUN fcis-7187 121 12 , , PUNCT fcis-7187 121 13 et et PROPN fcis-7187 121 14 al al PROPN fcis-7187 121 15 . . PUNCT fcis-7187 121 16 domain domain NOUN fcis-7187 121 17 - - PUNCT fcis-7187 121 18 adversarial adversarial ADJ fcis-7187 121 19 training training NOUN fcis-7187 121 20 of of ADP fcis-7187 121 21 neural neural ADJ fcis-7187 121 22 networks[j networks[j PROPN fcis-7187 121 23 ] ] X fcis-7187 121 24 . . PUNCT fcis-7187 122 1 the the DET fcis-7187 122 2 journal journal NOUN fcis-7187 122 3 of of ADP fcis-7187 122 4 machine machine NOUN fcis-7187 122 5 learning learn VERB fcis-7187 122 6 research research NOUN fcis-7187 122 7 , , PUNCT fcis-7187 122 8 2016 2016 NUM fcis-7187 122 9 , , PUNCT fcis-7187 122 10 17(1 17(1 NUM fcis-7187 122 11 ): ): PUNCT fcis-7187 122 12 2096 2096 NUM fcis-7187 122 13 - - SYM fcis-7187 122 14 2030 2030 NUM fcis-7187 122 15 . . PUNCT fcis-7187 123 1 [ [ X fcis-7187 123 2 5 5 X fcis-7187 123 3 ] ] X fcis-7187 123 4 chen chen PROPN fcis-7187 123 5 y y PROPN fcis-7187 123 6 , , PUNCT fcis-7187 123 7 li li PROPN fcis-7187 123 8 w w PROPN fcis-7187 123 9 , , PUNCT fcis-7187 123 10 sakaridis sakaridis PROPN fcis-7187 123 11 c c VERB fcis-7187 123 12 , , PUNCT fcis-7187 123 13 et et PROPN fcis-7187 123 14 al al PROPN fcis-7187 123 15 . . PROPN fcis-7187 123 16 domain domain PROPN fcis-7187 123 17 adaptive adaptive ADJ fcis-7187 123 18 faster fast ADV fcis-7187 123 19 r r NOUN fcis-7187 123 20 - - PUNCT fcis-7187 123 21 cnn cnn NOUN fcis-7187 123 22 for for ADP fcis-7187 123 23 object object NOUN fcis-7187 123 24 detection detection NOUN fcis-7187 123 25 in in ADP fcis-7187 123 26 the the DET fcis-7187 123 27 wild[c]//proceedings wild[c]//proceeding NOUN fcis-7187 123 28 of of ADP fcis-7187 123 29 the the DET fcis-7187 123 30 ieee ieee NOUN fcis-7187 123 31 conference conference NOUN fcis-7187 123 32 on on ADP fcis-7187 123 33 computer computer NOUN fcis-7187 123 34 vision vision NOUN fcis-7187 123 35 and and CCONJ fcis-7187 123 36 pattern pattern NOUN fcis-7187 123 37 recognition recognition NOUN fcis-7187 123 38 . . PUNCT fcis-7187 124 1 2018 2018 NUM fcis-7187 124 2 : : PUNCT fcis-7187 124 3 3339 3339 NUM fcis-7187 124 4 - - SYM fcis-7187 124 5 3348 3348 NUM fcis-7187 124 6 . . PUNCT fcis-7187 125 1 [ [ X fcis-7187 125 2 6 6 NUM fcis-7187 125 3 ] ] PUNCT fcis-7187 125 4 hsu hsu PROPN fcis-7187 125 5 h h PROPN fcis-7187 125 6 k k PROPN fcis-7187 125 7 , , PUNCT fcis-7187 125 8 yao yao PROPN fcis-7187 125 9 c c PROPN fcis-7187 125 10 h h PROPN fcis-7187 125 11 , , PUNCT fcis-7187 125 12 tsai tsai PROPN fcis-7187 125 13 y y PROPN fcis-7187 125 14 h h PROPN fcis-7187 125 15 , , PUNCT fcis-7187 125 16 et et PROPN fcis-7187 125 17 al al PROPN fcis-7187 125 18 . . PROPN fcis-7187 125 19 progressive progressive ADJ fcis-7187 125 20 domain domain NOUN fcis-7187 125 21 adaptation adaptation NOUN fcis-7187 125 22 for for ADP fcis-7187 125 23 object object NOUN fcis-7187 125 24 detection[c]//proceedings detection[c]//proceeding NOUN fcis-7187 125 25 of of ADP fcis-7187 125 26 the the DET fcis-7187 125 27 ieee ieee NOUN fcis-7187 125 28 / / SYM fcis-7187 125 29 cvf cvf NOUN fcis-7187 125 30 winter winter NOUN fcis-7187 125 31 conference conference NOUN fcis-7187 125 32 on on ADP fcis-7187 125 33 applications application NOUN fcis-7187 125 34 of of ADP fcis-7187 125 35 computer computer NOUN fcis-7187 125 36 vision vision NOUN fcis-7187 125 37 . . PUNCT fcis-7187 126 1 2020 2020 NUM fcis-7187 126 2 : : PUNCT fcis-7187 127 1 749 749 NUM fcis-7187 127 2 - - SYM fcis-7187 127 3 757 757 NOUN fcis-7187 127 4 . . PUNCT fcis-7187 128 1 [ [ X fcis-7187 128 2 7 7 X fcis-7187 128 3 ] ] X fcis-7187 128 4 tseng tseng PROPN fcis-7187 128 5 h h PROPN fcis-7187 128 6 y y PROPN fcis-7187 128 7 , , PUNCT fcis-7187 128 8 lee lee PROPN fcis-7187 128 9 h h PROPN fcis-7187 128 10 y y PROPN fcis-7187 128 11 , , PUNCT fcis-7187 128 12 huang huang PROPN fcis-7187 128 13 j j PROPN fcis-7187 128 14 b b PROPN fcis-7187 128 15 , , PUNCT fcis-7187 128 16 et et PROPN fcis-7187 128 17 al al PROPN fcis-7187 128 18 . . PUNCT fcis-7187 129 1 cross cross ADJ fcis-7187 129 2 - - ADJ fcis-7187 129 3 domain domain ADJ fcis-7187 129 4 few few ADJ fcis-7187 129 5 - - PUNCT fcis-7187 129 6 shot shot NOUN fcis-7187 129 7 classification classification NOUN fcis-7187 129 8 via via ADP fcis-7187 129 9 learned learn VERB fcis-7187 129 10 feature feature NOUN fcis-7187 129 11 - - PUNCT fcis-7187 129 12 wise wise ADJ fcis-7187 129 13 transformation[j transformation[j NOUN fcis-7187 129 14 ] ] PUNCT fcis-7187 129 15 . . PUNCT fcis-7187 130 1 arxiv arxiv PROPN fcis-7187 130 2 preprint preprint PROPN fcis-7187 130 3 arxiv:2001.08735 arxiv:2001.08735 NOUN fcis-7187 130 4 , , PUNCT fcis-7187 130 5 2020 2020 NUM fcis-7187 130 6 . . PUNCT fcis-7187 131 1 [ [ X fcis-7187 131 2 8 8 NUM fcis-7187 131 3 ] ] PUNCT fcis-7187 131 4 marino marino PROPN fcis-7187 131 5 s s PROPN fcis-7187 131 6 , , PUNCT fcis-7187 131 7 beauseroy beauseroy ADJ fcis-7187 131 8 p p X fcis-7187 131 9 , , PUNCT fcis-7187 131 10 smolarz smolarz ADJ fcis-7187 131 11 a. a. NOUN fcis-7187 131 12 unsupervised unsupervise VERB fcis-7187 131 13 adversarial adversarial ADJ fcis-7187 131 14 deep deep ADJ fcis-7187 131 15 domain domain NOUN fcis-7187 131 16 adaptation adaptation NOUN fcis-7187 131 17 method method NOUN fcis-7187 131 18 for for ADP fcis-7187 131 19 potato potato NOUN fcis-7187 131 20 defects defect NOUN fcis-7187 131 21 classification[j classification[j NOUN fcis-7187 131 22 ] ] PUNCT fcis-7187 131 23 . . PUNCT fcis-7187 132 1 computers computer NOUN fcis-7187 132 2 and and CCONJ fcis-7187 132 3 electronics electronic NOUN fcis-7187 132 4 in in ADP fcis-7187 132 5 agriculture agriculture NOUN fcis-7187 132 6 , , PUNCT fcis-7187 132 7 2020 2020 NUM fcis-7187 132 8 , , PUNCT fcis-7187 132 9 174 174 NUM fcis-7187 132 10 : : PUNCT fcis-7187 132 11 105501 105501 NUM fcis-7187 132 12 . . PUNCT fcis-7187 133 1 [ [ X fcis-7187 133 2 9 9 NUM fcis-7187 133 3 ] ] X fcis-7187 133 4 ganin ganin PROPN fcis-7187 133 5 y y PROPN fcis-7187 133 6 , , PUNCT fcis-7187 133 7 lempitsky lempitsky NOUN fcis-7187 133 8 v. v. ADP fcis-7187 133 9 unsupervised unsupervised ADJ fcis-7187 133 10 domain domain NOUN fcis-7187 133 11 adaptation adaptation NOUN fcis-7187 133 12 by by ADP fcis-7187 133 13 backpropagation[c]//international backpropagation[c]//international ADJ fcis-7187 133 14 conference conference NOUN fcis-7187 133 15 on on ADP fcis-7187 133 16 machine machine NOUN fcis-7187 133 17 learning learning NOUN fcis-7187 133 18 . . PUNCT fcis-7187 133 19 pmlr pmlr NOUN fcis-7187 133 20 , , PUNCT fcis-7187 133 21 2015 2015 NUM fcis-7187 133 22 : : PUNCT fcis-7187 133 23 1180 1180 NUM fcis-7187 133 24 - - SYM fcis-7187 133 25 1189 1189 NUM fcis-7187 133 26 . . PUNCT