Nkeji Ihe Nlereanya Ọ bụla
Ihe Nlereanya Nkebi Ihe Ọ bụla (SAM) bụ Meta Ihe nlere ntọala AI maka nkewa onyonyo: nyere isi okwu, igbe ma ọ bụ akara siri ike, ọ na-akọwapụta ihe kwekọrọ ozugbo.
Nchịkọta
It was built to generalize to objects and images it never saw during training, making segmentation a promptable task.
Ime miri emi
N'ịbụ nke Meta AI weputara na 2023, SAM reframes ngalaba dị ka nsogbu nwere ike ime ngwa ngwa: ị na-enye ya ngwa ngwa (pịa, igbe, ihe nkpuchi, ma ọ bụ akara sitere na ederede) wee weghachi otu ihe ma ọ bụ karịa. Ike ya sitere na nha nha: a zụrụ ya na SA-1B, ihe ndekọ data nke ihe karịrị ijeri 1 ijeri ihe onyonyo nde iri na otu, ejiri igwe nkọwa ihe nlegharị anya na-arụ. N'usoro ihe owuwu, SAM nwere ihe ngbanwe onyonyo dị arọ na-agba ọsọ otu ugboro n'otu onyonyo ọ bụla, ihe ngbanwe ngwa ngwa ngwa ngwa, yana ihe nkpuchi ngwa ngwa, yabụ enwere ike iweghachi otu onyonyo agbakwunyere na mmekọrịta ozugbo. Ọ na-enye ohere ịnyefe efu-shot gaa n'ọtụtụ ọrụ. SAM 2, ewepụtara na 2024, gbatịa nke a na vidiyo, na-enyocha ihe n'ofe okpokolo agba.
Nghọta nka nka
SAM na-eji ihe ngbanwe onyonyo (ViT) ihe ngbanwe, nke a na-azụkarị ya na nkpuchi autoencoding, iji mepụta nnukwu ihe ntinye onyonyo. A na-edobe ngwa ngwa n'ime tokens, yana ihe ngbanwe dabere na mgbanwe mgbanwe nke nwere nlebara anya nlebara anya na-eme ka akara ngosi nwere ihe onyonyo a na-etinye na ihe mkpuchi gbakwunyere akara ntụkwasị obi. Iji dozie nghọtahie (ịpịpị nwere ike ịpụta bọtịnụ, uwe elu, ma ọ bụ mmadụ), SAM na-ebu amụma ọtụtụ masks dị mma n'otu oge wee wepụta ha, na-ahapụ iji ala ala ma ọ bụ ihe ọzọ kpalie agbagha.
Mmetụta atụmatụ
Ọsọ na ọnụ ọgụgụ
Visual AI nwere ike megharịa nyocha, nchọpụta na mkpado ọrụ n'ọtụtụ.
Mee nhọrọ
Otu ndị na-emepụta ihe nwere ike imepụta echiche ngwa ngwa site na ngbanwe akwụkwọ ntuziaka ole na ole.
Team na usoro ọrụ
Ọrụ nwere ike iji onyonyo na akara vidiyo siri ike ịhazi.
Ọdịnihu nke Nkebi Ihe Nlereanya Ọ bụla
SAM abụrụla ọkpụkpụ azụ ndabere maka ngwaọrụ nkọwa, onyonyo ahụike, robotics na pipeline AR, na-ejikọkarị ya na ihe nchọpụta ma ọ bụ ụdị ederede maka 'akụkụ aha' nke mepere emepe. Na-atụ anya ụdị dị mfe, ngwa ngwa dị iche iche (MobileSAM, EfficientSAM) maka iji ngwaọrụ, njikọta miri emi na asụsụ maka nkewa ederede zuru oke, yana ịgbasawanye na vidiyo na 3D. Dị ka ihe nlereanya ntọala, a na-ejikarị ihe ntinye ya eme ihe dị ka oyi akwa nghọta na-enye usoro ndị ọzọ nri.
Mmejuputa n'ezie n'ụwa
Usoro nkọwa ihe onyonyo na-eji SAM mee ka ndị na-ede aha pịa otu ugboro wee mepụta ihe mkpuchi ihe kpọmkwem, na-egbutu oge ịkpọ aha.
Ndị nchọpụta na-emegharị SAM (dịka, MedSAM) iji depụta akụkụ ahụ na etuto ahụ na nyocha CT na MRI.
Ndị na-edezi foto na vidiyo na-ejikọta SAM iji bepụ isiokwu ma ọ bụ wepụ ndabere na otu ọpịpị.
SAM 2 egwu na akụkụ ihe n'ofe vidiyo vidiyo maka mmetụta AR na nghọta robotics.
Ihe ize ndụ & okporo ụzọ nche
Ikike onyonyo na nkwenye nwere ike bụrụ ihe egwu dị n'iwu ma ọ bụrụ na edoghị anya.
Ọrụ nlereanya nwere ike ịdịgasị iche n'ofe ọkụ, igwe mmadụ, na gburugburu.
Enwere ike ghara ịhụ ihe dị mma ma ọ bụrụ na enyochaghị oke ntụkwasị obi.
Map mmejuputa
Kọwaa ụkpụrụ nnabata maka nkenke, icheta, na ụgwọ njehie.
Nwalee na data dabara na ọnọdụ mmepụta n'ezie.
Tinye nyocha mmadụ maka obere obi ike ma ọ bụ amụma mmetụta dị elu.
Sochie ihe nlere anya wee megharịa ka emechara mgbanwe igwefoto ma ọ bụ dataset.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Segment Anything Model quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
Ụdị agbanwe agbanwe
Ajụjụ a na-ajụkarị
What is Segment Anything Model?
Ihe Nlereanya Nkebi Ihe Ọ bụla (SAM) bụ Meta Ihe nlere ntọala AI maka nkewa onyonyo: nyere isi okwu, igbe ma ọ bụ akara siri ike, ọ na-akọwapụta ihe kwekọrọ ozugbo. Ewubere ya ka ọ chịkọta ihe na ihe onyonyo ọ na-ahụtụbeghị n'oge ọzụzụ, na-eme ka nkewa bụrụ ọrụ a na-eme ngwa ngwa.
Kedu isi echiche na-eme SAM ka ọ bụrụ 'ihe atụ ntọala' maka nkewa?
SAM na-emegharị akụkụ dị ka nke a na-eme ngwa ngwa ma emebere ya maka ịnyefe ihe efu na ihe oyiyi na-abụghị usoro ọzụzụ ya.
Ihe dị ka ole ka a na-eji na-azụ SAM dataset SA-1B?
SA-1B nwere ihe mkpuchi ihe karịrị ijeri 1 n'ihe onyonyo nde iri na otu, ejiri igwe nkọwa ihe nlegharị anya arụrụ.
Kedu ihe kpatara SAM ji kewaa ihe owuwu ya ka ọ bụrụ ihe ngbanwe onyonyo dị arọ yana ihe ngbanwe/dekoder ngwa ngwa ngwa ngwa?
Ihe onyonyo dị oke ọnụ na-agba ọsọ otu ugboro; mgbe ahụ, ngbanwe ngwa ngwa ngwa ngwa na nkpuchi nkpuchi na-enye ohere ịmegharị ngwa ngwa na mmekọrịta nke otu onyonyo ahụ.
Kedu ka SAM si ejikwa ngwa ngwa na-enweghị isi, dị ka otu ọpịpị nke nwere ike ịpụta ọtụtụ ihe?
SAM na-ewepụta ọtụtụ ihe nkpuchi n'ọkwa nwere akara ka enwere ike idozi enweghị isi site na ogo ma ọ bụ mkpali ọzọ.
Kedu ụdị ọkpụkpụ azụ ka SAM na-eji tinye koodu ntinye onyonyo?
Ihe ngbanwe onyonyo nke SAM bụ ihe ntụgharị ọhụụ, nke a na-azụkarị ya site na nkpuchi autoencoding, na-ewepụta nnukwu ihe onyonyo.