Ntuziaka Visual AI

Omimi Ihe ọ bụla Omimi Monocular

DepthAnything bụ ụdị ntọala nke na-eme atụmatụ ka pikselụ ọ bụla dị anya site na otu foto nkịtị, na-enweghị ngwaike pụrụ iche.

2 nkeji na-agụEmelitere ikpeazụ

Nchịkọta

It made robust, general-purpose depth sensing cheap and accessible for anything from phones to robots.

Ime miri emi

DepthAnything (2024, nke ndị nyocha wepụtara gụnyere ndị nọ na TikTok/ByteDance na HKU) na-eme atụmatụ omimi nke monocular: ịkọ maapụ omimi site na otu onyonyo RGB. Ọganihu ya bụ ọnụ ọgụgụ: kama ịdabere naanị na ntakịrị data omimi akpọrọ dị, ndị otu ahụ wuru injin nke ejiri akara ngosi nde 62 na-enweghị akara site na iji ụdị onye nkuzi wee zụọ nwa akwụkwọ na nnukwu ụlọ ọrụ a. Nke a na-enye nkọwapụta efu efu n'ofe ime ụlọ, n'èzí na ihe nkiri pụrụ iche. Nsonaazụ mbụ a na-egosi omimi dị omimi (nke pikselụ dị nso ma ọ bụ karịa, ọ bụghị kpọmkwem mita). Omimi ihe ọ bụla V2 (n'etiti 2024) na-eme ka nkọwa dị mma site n'ịzụ onye nkuzi banyere data sịntetik na ezi-okwu zuru oke, wee na-atụgharị na foto n'ezie, na-edozi akụkụ ndị na-adịghị mma na njehie ihe doro anya.

Nghọta nka nka

Ọ na-eji DINov2 ntụgharị ihe ntụgharị ọhụụ na-enye isi amụma oke ụdị DPT. Isi aghụghọ bụ distillation nke a na-ahụ maka ọkara: onye nkuzi zụrụ azụ na data pseudo-labeled ọtụtụ nde onyonyo enweghị akara, nwa akwụkwọ na-amụtakwa na ha abụọ. V2 na-agbanwe ezigbo akara mkpọtụ maka data sịntetik na omimi pikselụ zuru oke, wee tụgharịa azụ na foto n'ezie, na-ewepụ ụkọ na mkpọtụ nkọwa nkọwa dị omimi ma na-edobe oke.

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 Omimi Ihe ọ bụla Omimi Monocular

Na-atụ anya ntinye siri ike na enyo AR, igwefoto ama, na robotics ebe LiDAR raara onwe ya nye dị oke ọnụ ma ọ bụ buru ibu. Ụdị metric dị iche iche nke na-emepụta mita n'ezie, gbakwunyere ụdị vidiyo nwere ịdị omimi nwa oge (enweghị flicker n'etiti okpokolo agba), na-aga ngwa ngwa. Ka ụdịdị ndị a na-ebelata ka ọ na-agba ọsọ na ngwaọrụ ozugbo, nghọta 3D otu igwefoto ga-abụ ikike ndabara, ịgbakọ mkpokọta ohere, igodo kwụụrụ onwe ya, yana nrụpụta ihe ngosi 3D.

Mmejuputa n'ezie n'ụwa

Na-emepụta maapụ omimi iji chụpụ ezigbo nhụsianya ndabere (bokeh) na foto eserese ekwentị nwere otu oghere.

Inye nghọta ihe mgbochi 3D maka drones dị ọnụ ala na robots na-enweghị LiDAR ma ọ bụ igwefoto stereo.

Ịmepụta maapụ ihe ndozi omimi maka ControlNet ka ndị na-emepụta ihe onyonyo wee chekwaa geometry ọnọdụ.

Na-atụgharị foto na ihe nkiri 2D ka ọ bụrụ mmetụta 3D ma ọ bụ parallax maka ihe ngosi VR na stereoscopic.

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 DepthAnything Monocular Depth 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

Ntuziaka na-esote

Ntụle omimi nke monocular

Ajụjụ a na-ajụkarị

What is DepthAnything Monocular Depth?

DepthAnything bụ ụdị ntọala nke na-eme atụmatụ ka pikselụ ọ bụla dị anya site na otu foto nkịtị, na-enweghị ngwaike pụrụ iche. Ọ mere omimi siri ike, ebumnuche izugbe na-ahụta dị ọnụ ala yana ịnweta maka ihe ọ bụla site na ekwentị ruo na robots.

Kedu ihe ntụle omimi 'monocular' pụtara?

Monocular pụtara omimi ka a na-enweta site na otu onyonyo igwefoto (mono), na-enweghị ụzọ stereo ma ọ bụ ihe mmetụta na-arụ ọrụ.

Kedu ihe bụ isi atụmatụ DepthAnything maka imezu mkpokọta siri ike?

Ndị otu ahụ wuru injin data nke depụtara puku kwuru iri nde foto ndị a na-edeghị aha, na-agbasa n'ọ̀tụ̀tụ̀ ọzụzụ n'ụzọ dị egwu.

Kedu ihe ngbanwe ọkpụkpụ azụ ka DepthAnything na-ewuli na?

DepthAnything na-eji koodu DINOv2 ViT jikọtara ya na isi amụma ogologo ụdị DPT.

Kedu ụdị omimi nke mbụ DepthAnything na-ebute ụzọ?

Ụdị ntọala ahụ na-ebu amụma ịtụ ịtụ ihe omimi karịa karịa ogologo metrik zuru oke.

Kedu ka DepthAnything V2 siri meziwanye nkọwa na akụkụ dị mma?

V2 zụrụ onye nkuzi na onyonyo sịntetịt na omimi zuru oke, wee tụgharịa gaa na foto n'ezie maka oke dị nkọ.