Overview
DETR (DEtection TRAnsformer) inogadziridza kutariswa kwechinhu seyakananga set-prediction dambudziko rinogadziriswa neshanduko, kubvisa nhanho dzakagadzirwa nemaoko senge mabhokisi eanchor uye isiri-yakanyanya kudzvinyirira. Izvo zvine basa nekuti zvakapa kuwonekwa kwakachena, kumagumo-kusvika-kumagumo pombi yakafuridzira mafungu eshanduro-yakavakirwa muono modhi.
DETR Transformer Detection ndeyekombuta-yekuona workflows inodudzira kana kugadzira midhiya yekuona yekuongorora, mashandiro, uye kugadzira.
Deep Dive
Yakaunzwa neFacebook AI muna 2020, DETR inosanganisa CNN musana neshanduro encoder-decoder. Iyo CNN inobvisa maficha emifananidzo; iyo encoder inosanganisa mamiriro epasi rose pamufananidzo wese; uye decoder inotora yakagadziriswa seti yezvakadzidzwa 'mivhunzo yechinhu' uye inoshandura imwe neimwe kuita chinhu chaonekwa (kirasi pamwe nebhokisi rinosungirirwa) kana 'hapana chinhu' mhedzisiro. Chinhu chakanyanya kukosha ndeye bipartite matching: panguva yekudzidziswa algorithm yeHungary inowana imwe-kune-imwe basa pakati pekufungidzira uye pasi-chokwadi zvinhu, saka modhi inodzidza kuburitsa rakasiyana bhokisi pachinhu chakananga. Izvi zvinobvisa kusanyanya kudzvinyirira uye anchor tuning. Kutengeserana kwacho kwainonoka kuchinjika uye kushomeka kwechinhu chidiki-chinhu, icho chinotevera senge Deformable DETR yakagadziriswa.
Technical Insight
DETR's defining mechanism ndiyo yekuseti-yakavakirwa kurasikirwa neHungarian kuenzanisa. Panzvimbo pekunwisa zviuru zvemabhokisi eanchor, inoburitsa nhamba yakatarwa yekufungidzira (kazhinji 100 chinhu mibvunzo) uye inoafananidza imwe-kune-imwe kune yechokwadi zvinhu, kuranga zvese zviri zviviri kupatsanurwa nebhokisi kukanganisa pamapeya akafananidzwa uye kusunda mibvunzo isingaenzaniswi kune 'hapana chinhu.' Nekuti kuenzanisa ndeye-kune-imwe, kudzokororwa kunoonekwa kunodzvanywa nedhizaini kwete neyakasiyana post-yekugadzirisa nhanho.
Mastering DETR Transformer Detection
DETR (DEtection TRAnsformer) inogadziridza kutariswa kwechinhu seyakananga set-prediction dambudziko rinogadziriswa neshanduko, kubvisa nhanho dzakagadzirwa nemaoko senge mabhokisi eanchor uye isiri-yakanyanya kudzvinyirira. Izvo zvine basa nekuti zvakapa kuona kwakachena, kumagumo-kusvika-kumagumo pombi iyo yakafuridzira mafungu eshanduro-yakavakirwa-muono modhi. DETR Transformer Detection ndeyekombuta-yekuona workflows inodudzira kana kugadzira midhiya yekuona yekuongorora, mashandiro, uye kugadzira. Kuti uvake kunzwisisa kwakadzama, bata DETR Transformer Detection semuenzaniso wekushandisa, kwete chinhu chimwe chete: tsanangura zvinodiwa, kujekesa fungidziro, uye patsanura zvinogona kuitwa nehurongwa hwakavimbika kubva kune zvichiri kuda kutonga kwenyanzvi.
Mukuita, zvikwata zvakasimba zvinoshandisa DETR Transformer Detection chiyero chechokwadi nezvinhu zvekushanda semhando yedata, kusiyana kwemwenje, uye kuenderana kwemazita. Ivo vanonyora zvakajeka maitiro ebudiriro, bvunzo vachipokana ne data rechokwadi uye mafambiro ebasa, uye iterate zvichibva pane zvakacherechedzwa maitiro ekutadza kwete kuhwina-nguva imwe chete yebhenji. Apa ndipo apo kunzwisisa kwe theoretical kunoshanduka kuve kugona kwakasimba pane chigadzirwa, mutemo, uye mashandiro.
Visual AI inogona kuita otomatiki yekuongorora, yekuona, uye yekumaka mabasa pachiyero. Panguva imwecheteyo, kodzero dzeMufananidzo uye kubvumirwa kunogona kuve njodzi dzepamutemo kana hunhu husina kujeka. Nzira yakatsiga ndeyekubatanidza kukurumidza kuyedza nekutonga: mhanyisa vatyairi vendege, tora humbowo, buritsa matanda esarudzo, uye urambe uchivandudza chengetedzo semaitiro emuenzaniso, zvinotarisirwa nemushandisi, uye zvinodikanwa zvekutonga.
Strategic Impact
Visual AI inogona kuita otomatiki yekuongorora, yekuona, uye yekumaka mabasa pachiyero.
Visual AI inogona kuita otomatiki yekuongorora, yekuona, uye yekumaka mabasa pachiyero. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Zvikwata zvekugadzira zvinogona prototype pfungwa nekukurumidza nekudzokororwa kwemaoko mashoma.
Zvikwata zvekugadzira zvinogona prototype pfungwa nekukurumidza nekudzokororwa kwemaoko mashoma. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Mashandisirwo anogona kushandisa masaini emifananidzo nemavhidhiyo ayo aimbove akaoma kugadzirisa.
Mashandisirwo anogona kushandisa masaini emifananidzo nemavhidhiyo ayo aimbove akaoma kugadzirisa. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Real-World Implementation
Kutsvaga uye tsiva vanofamba netsoka uye mota mune yakazvimirira-kutyaira yekutsvagisa dataset
Kusimbisa panoptic segmentation kana yawedzerwa kune-pixel mask kufanotaura
Kushanda semusana wezvivakwa zveakavhurika-mazwi uye zvigadziriso zvepasi
Kutsvaga zvinhu mumasherufu ezvitoro mapikicha pasina kugadzirisa saizi dzeanchor pane dataset
Maitiro Ekuita
DETR Transformer Kuonekwa mukuita
Kutsvaga uye tsiva vanofamba netsoka uye mota mune yakazvimirira-kutyaira yekutsvagisa dataset.
Kutsvaga uye kurova vafambi netsoka uye mota mune yakazvimirira-kutyaira yekutsvagisa datasets Zvikwata zvinowanzowana mhedzisiro iri nani kana vatsanangura hunhu hwepamberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa nemitengo yekukanganisa nekufamba kwenguva.
DETR Transformer Kuonekwa mukuita
Kusimbisa panoptic segmentation kana yawedzerwa kune-pixel mask kufanotaura.
Powering panoptic segmentation kana yakawedzerwa kune per-pixel mask kufanotaura Zvikwata zvinowanzowana mhedzisiro irinani kana vachitsanangudza mhando kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
DETR Transformer Kuonekwa mukuita
Kushanda semusana wezvivakwa zveakavhurika-mazwi uye zvigadziriso zvepasi.
Kushanda semhare yekuvaka yeakavhurika-mazwi uye ekuisa pasi matikiti Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa nemitengo yekukanganisa nekufamba kwenguva.
DETR Transformer Kuonekwa mukuita
Kutsvaga zvinhu mumasherufu ezvitoro mapikicha pasina kugadzirisa saizi dzeanchor pane dataset.
Kuwana zvinhu muzvitoro zvemasherufu mapikicha pasina kuchinjisa saizi yeanchor padhataset Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
Njodzi & Guardrails
Kodzero dzemifananidzo uye kubvumirwa kunogona kuve njodzi dzepamutemo kana provenance isina kujeka.
Kuita kwemuenzaniso kunogona kusiyanisa kupenya, huwandu hwevanhu, uye nharaunda.
Manyepo enhema anogona kusacherechedzwa kunze kwekunge zvikumbaridzo zvekuvimba zvikatariswa.
Implementation Roadmap
Tsanangura maitiro ekugamuchirwa echokwadi, kurangarira, uye mutengo wekukanganisa.
Tsanangura maitiro ekugamuchirwa echokwadi, kurangarira, uye mutengo wekukanganisa. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Edzai nedata rinoenderana nemamiriro chaiwo ekugadzira.
Edzai nedata rinoenderana nemamiriro chaiwo ekugadzira. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Wedzera ongororo yemunhu kune yakaderera-kusavimbika kana yakakwirira-inokanganisa kufanotaura.
Wedzera ongororo yemunhu kune yakaderera-kusavimbika kana yakakwirira-inokanganisa kufanotaura. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Tevera modhi kudonha uye simbisa mushure mekuchinja kwekamera kana dataset.
Tevera modhi kudonha uye simbisa mushure mekuchinja kwekamera kana dataset. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.