Umbhali: (1) I-Asaduz Zaman, i-Dept. ye-Data Science kunye ne-Artificial Intelligence, i-Faculty of Information Technology, i-Monash University, eAustralia (asaduzzaman@monash.edu); (2) Vanessa Kellermann, i-Department of Environment and Genetics, School of Agriculture, Biomedicine, and Environment, University of La Trobe, Australia (v.kellermann@latrobe.edu.au); (3) U-Alan Dorin, i-Dept. of Data Science and Artificial Intelligence, i-Faculty of Information Technology, i-Monash University, eAustralia (alan.dorin@monash.edu). Authors: (1) I-Asaduz Zaman, i-Dept. ye-Data Science kunye ne-Artificial Intelligence, i-Faculty of Information Technology, i-Monash University, eAustralia (asaduzzaman@monash.edu); (2) Vanessa Kellermann, i-Department of Environment and Genetics, School of Agriculture, Biomedicine, and Environment, University of La Trobe, Australia (v.kellermann@latrobe.edu.au); (3) U-Alan Dorin, i-Dept. of Data Science and Artificial Intelligence, i-Faculty of Information Technology, i-Monash University, eAustralia (alan.dorin@monash.edu). Umbala we-Left I-Abstract kunye ne-Introduction Iimpawu eziqhelekileyo Umgangatho Imiphumo kunye nokuxhumana Ukuhlaziywa kunye ne-references Ukucinga Uluhlu lo mveliso ibonise i-re-identification data ye-animal efanelekileyo, i-concept entsha kunye ne-technique esebenzayo yokufumana iziphumo ezidlulileyo ze-organisms kwi-archived data, leyo ibonise iindlela zokuzonwabisa ze-chronological re-identification ezivamile kwi-re-identification ye-behavioral research. Ukukhishwa kwe-individual key phakathi kweengxaki ezininzi kunokufumaneka ngexesha elide kwi-experiment ukuba ibonise nge-behavior behavior elidlulileyo emva kwexesha le-undifferentiated yokusebenza. Ngokuvamile, izifundo ze-longitudinal ziyafumaneka nokufumana ukuchithwa kwe-subject ngexperiment I-1 Introduction Kwi-behavioral studies longitudinal, ukucacisa ama-subjects ezimbalwa ngexesha, ukucacisa xa zibonakalisa kwakhona, kwaye kwakhona xa zibonakalisa kwimibelelwano ezilandelayo, kubalulekile yokufumana umzimba [2]. Re-identification (re-id) zezilwanyana ezincinane, ezifanayo ngokufanelekileyo, ezifana nezilwanyana, ingasetyenziswa ngama-markers yemvelo okanye i-tag [2, 4, 14, 15]. Nangona kunjalo, ezi zibonakalisa umzimba we-subjects [7]. I-re-id ngaphandle kwe-marker ingasetyenziswa ngempumelelo ukucacisa izilwanyana zonyango [16]. Nangona kunjalo, oku kuthatha izilwanyana ezifanelekileyo kakhulu, ezifana nezilwanyana, kwaye Iingxaki ze-morphological nge-slit and tear ingathintela iimpumelelo yokuguqula umntu. Ukuba i-subject ifunyenwe okanye i-visual altered ngexesha lokusebenza, iimveliso ezisetyenziswa kwi-training image classification software ukucacisa iya kufuneka zithunyelwe. Le inefficiency isetyenziswa ngexabiso kwizinto ezininzi ezininzi ngexabiso ukuba amancinci izihlangu ukuya ekugqibeleni, kwaye abafutshane iya kubonisa ubungakanani obungapheliyo obungapheliyo, ezifana ukufundisa isicelo, ukutshintsha isisombululo, okanye ukufumana isisombululo esifunyenweyo [8]. Ngoko ke, ukucaciswa kwakhona kwizinto eziphambili ukusuka kw -i-identification (i-retro-id) ukuyisombulula le ngxaki. iimveliso Ngaphandle kokufunda iimodeli zokufunda kwiimodeli yokuqala (i-day one) iinkcukacha kunye nokufunda iindidi ngexesha lwezobugcisa ngexesha lokusebenza, sincoma ukuba kunokuba kunzima kakhulu ukwenza ingxaki oluthile. Yeyona, ngexesha elinye kufuneka usebenzise i-algorithms zethu kwiimodeli zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo zeemvavanyo Thina ukubonisa ukuba imodeli olufunyenwe kwi-imeyile ye-insect data ukususela ngosuku lokuqala kwaye i-tested ngenxa yayo yokusebenza kwe-re-id insects ukuya ngosuku ye-N, iya kubonisa ukusebenza efanayo kunye imodeli olufunyenwe kwi-day ye-N data kwaye i-tested kwi-retro-id insects ukususela ngosuku lokuqala. Thina uyifunyenwe ngama-monitoring i-15 iintlobo ze-reed bees kwiintsuku ze-5 ezininzi. Lezi ama-pollinators ze-semi-social ziyafumaneka kakhulu kwi-phenotypic (i-Figure 1) kwaye ziyafumaneka ngokwemvelo ngexesha elinye, ngexesha elandelayo, kwenza i-re-id yobungcali kwim Oku kunokwenzeka kwi-archiv phantsi kolawulo lwe-CC BY 4.0 DEED. Oku kunokwenzeka kwi-archiv phantsi kolawulo lwe-CC BY 4.0 DEED. Zifumaneka kwi-Archiv
