Ntuziaka Visual AI

Nkwenye omume

Ịmata ihe omume bụ ọrụ ịkụziri kọmputa iji chọpụta ihe ndị mmadụ ma ọ bụ ihe * na-eme * na vidiyo - ịgba ọsọ, ifegharị, ịdaba, imeghe ọnụ ụzọ - ọ bụghị naanị ihe na-apụta n'otu etiti.

2 nkeji na-agụEmelitere ikpeazụ

Nchịkọta

It matters because understanding motion over time unlocks applications from sports analytics to elderly fall detection.

Ime miri emi

Nchọpụta ihe omume na-agabiga nhazi onyonyo kwụ ọtọ site n'ịtụgharị uche maka otu pikselụ si agbanwe n'ofe oge. Otu etiti nwere ike igosi mmadụ n'etiti ikuku; naanị usoro na-egosi ma ha na-awụli elu, na-ada ma ọ bụ na-egwu mmiri. Sistemu mmalite ejiri aka rụọ ọrụ ngagharị dị ka mgbagharị anya na nnukwu trajectories. Ụzọ nke oge a na-eji netwọk miri emi: usoro nhazi nke abụọ na-emepụta ihe (RGB frames) na mmegharị (ntụgharị anya) iche iche; Netwọk convolutional 3D (dị ka C3D na I3D) nzacha slide site na oghere * na oge; na ihe ntụgharị vidiyo (TimeSformer, VideoMAE) tinye nlebara anya n'ofe patches spatio- temporal patches. Ọkọlọtọ akara gụnyere Kinetics (klaasị omume mmadụ 700 sitere na YouTube), UCF101, na Ihe-Ihe, nke na-amanye ụdịdị ịghọta ntụzịaka oge kama ịbụ naanị ọnọdụ ọnọdụ.

Nghọta nka nka

Isi ihe ịma aka bụ imezi akụkụ nke oge. Mgbagha 3D na-agbatị nzacha 2D nkịtị yana axis dị omimi na-agbasa ọtụtụ okpokolo agba, yabụ ọ na-amụta usoro ngagharị ozugbo. Aghụghọ I3D 'na-ebuli ibu' site na netwọk onyonyo 2D emebere ya na ImageNet n'ime 3D site na ịmegharị ha n'ofe oge, na-enye mmalite siri ike. Ụzọ iyi abụọ kama na-enyefe ngwa anya agbagoro agbakọọ n'ime ngalaba dị iche, na-edobe mmegharị ahụ nke ọma wee jikọta ya na njirimara ọdịdị.

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 Nghọta Omume

Ugbo a na-atụgharị gaa na ndị na-agbanwe vidiyo na-arụ ọrụ nke ọma na ọzụzụ ọzụzụ onwe onye na-ahụ maka ya (ihe ngosi vidiyo masked) nke na-amụta site na eserese na-enweghị aha, na-ebelata ịdabere na nkọwa dị oke ọnụ. Na-atụ anya njikọta siri ike na ụdị asụsụ multimodal ka sistemu nwere ike ọ bụghị naanị ịkpọ omume mana kọwaa ma tụgharịa uche gbasara ha n'asụsụ eke. Nkwenye n'oge na ngwaọrụ maka wearables, robotics, na igwefoto smart bụ oke oke, n'akụkụ ọmarịcha mara mma nke na-amata ọdịiche dị nro na nso nso a.

Mmejuputa n'ezie n'ụwa

Usoro nchọpụta ọdịda n'ụlọ ndị agadi na-eme ka ndị ọrụ mara mgbe onye bi na ya dara, na-amata ọdịiche dị n'ọdịda na ịnọdụ ala ma ọ bụ dinara ala.

Usoro nyocha egwuregwu na-akpado ozi na-akpaghị aka, ihe mgbakasị ahụ, na gbaa n'ihe nkiri egwuregwu maka nkuzi na ihe ngosi mgbasa ozi.

Nleba anya na nchekwa nchekwa nke na-egosipụta omume na-adịghị mma dị ka ọgụ, ịkwa iko, ma ọ bụ onye na-arịgoro ngere.

Ntụgharị na-achịkwa mmegharị ahụ yana ngwa ahụike na-agụta reps na ịlele ụdị mmega ahụ site n'ịmata mmegharị ahụ ka oge na-aga.

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

1

Kọwaa ụkpụrụ nnabata maka nkenke, icheta, na ụgwọ njehie.

2

Nwalee na data dabara na ọnọdụ mmepụta n'ezie.

3

Tinye nyocha mmadụ maka obere obi ike ma ọ bụ amụma mmetụta dị elu.

4

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 Action Recognition quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Malite ajụjụ

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Ajụjụ a na-ajụkarị

What is Action Recognition?

Ịmata ihe omume bụ ọrụ ịkụziri kọmputa iji chọpụta ihe ndị mmadụ ma ọ bụ ihe * na-eme * na vidiyo - ịgba ọsọ, ifegharị, ịdaba, imeghe ọnụ ụzọ - ọ bụghị naanị ihe na-apụta n'otu etiti. Ọ dị mkpa n'ihi na ịghọta mmegharị n'ime oge na-emeghe ngwa sitere na nyocha egwuregwu ruo nchọpụta ọdịda ndị agadi.

Kedu ihe dị iche n'ụzọ doro anya njirimara omume na nhazi onyonyo nkịtị?

Ụdị njirimara omume na-agagharị n'usoro nke okpokolo agba, ebe ọ bụ na otu onyonyo na-adịkarị enweghị ike ikpughe ma mmadụ ọ na-awụ elu, na-ada ma ọ bụ na-egwu mmiri.

N'ime netwọkụ ama ama ihe omume iyi abụọ, gịnị ka iyi nke abụọ na-ahazi?

Ihe owuwu iyi abụọ na-eji otu alaka maka mpụta (Freemu RGB) yana ngalaba nke abụọ maka ịgbasa ngwa anya, nke na-egosipụta n'ụzọ doro anya mmegharị n'etiti okpokolo agba.

Kedu ihe convolution 3D na-agbakwunye ma e jiri ya tụnyere ọkọlọtọ 2D convolution?

Ihe ngbanwe 3D na-agbatị nzacha ahụ na nha omimi/oge, na-ahapụ ya ka ọ mụta usoro ngagharị ozugbo n'ofe okpokolo agba na-esochi.

Gịnị bụ aghụghọ 'ọnụahịa' nke ụdị I3D na-eji?

I3D 'na-ebuli ibu' a zụrụ azụ na onyonyo 2D (dị ka ImageNet) n'ime 3D site na iṅomi ha n'akụkụ axis nwa oge, na-enye mmalite siri ike.

Kedu ihe kpatara ihe data ihe-Ihe-Something ji ama ama maka njirimara mmemme?

Emebere ihe-Ihe mere ka omume ahụ dabere n'usoro oge na mmegharị ahụ, na-egbochi ụdị aghụghọ site na iji ndabere ma ọ bụ ọdịdị ihe naanị.