Mip-NeRF ati Awọn aaye Radiance Anti-Aliased
Mip-NeRF ṣe atunṣe blurry, awọn ohun-ọṣọ ti o ni jagidi ti o kọlu NeRF atilẹba nigbati o ba ṣe awọn iwoye ni awọn ọna jijin tabi awọn ipinnu.
Akopọ
O ṣe eyi nipa titele awọn cones dipo awọn egungun tinrin ti ko ni ailopin, ṣiṣe oju iṣẹlẹ 3D ti o mu ki o didasilẹ ati yiyara lati ṣe ikẹkọ.
Jin Dive
NeRF atilẹba ṣe afihan iṣẹlẹ kan pẹlu awọn egungun tinrin, aaye kan ni akoko kan, ati ifunni ipo 3D kọọkan sinu nẹtiwọọki nkankikan. Iṣoro naa: aaye kan kọju iye ti iṣẹlẹ ti ẹbun kan bo nitootọ. Piksẹli nitosi kamẹra n rii agbegbe kekere kan; ẹbun kanna ti o jinna ri nla kan. Ṣiṣayẹwo wọn ni aami kan nfa aliasing - fifẹ ati jaggies bi o ṣe sun-un tabi gbe. Mip-NeRF (Barron et al., 2021) rọpo ray kọọkan pẹlu konu kan ati pe o pin si awọn frustums conical. Dipo fifi koodu si aaye kan, o ṣe koodu agbegbe inu inu ibanujẹ kọọkan nipa lilo fifi koodu ipo isọpọ (IPE), isunmọ iwọn didun pẹlu Gaussian kan. Eyi jẹ ki nẹtiwọọki multiscale kan ṣe ipinnu eyikeyi ni mimọ, gige aṣiṣe ati akoko ikẹkọ ni pataki.
Imọ-imọ-ẹrọ
Ẹtan bọtini naa jẹ ifidipo ipo ipo. Standard NeRF maapu aaye kan nipasẹ ese ati awọn iṣẹ cosine ni ọpọlọpọ awọn loorekoore. Mip-NeRF dipo isunmọ conical frustum bi Gaussian pupọ ati ṣe iṣiro iye ti a nireti ti awọn sinusoids yẹn lori Gaussian yẹn. Awọn ẹya ara ẹrọ igbohunsafẹfẹ giga ti o yatọ pupọ ni inu frustum nla kan gba aifọwọyi laifọwọyi si odo, titi di isisiyi tabi awọn agbegbe isokuso lo alaye iduroṣinṣin kekere-igbohunsafẹfẹ - deede ihuwasi atako-aliasing ti awọn mipmaps ni awọn aworan Ayebaye.
Ipa Ilana
Iyara ati iwọn
Visual AI le ṣe adaṣe adaṣe, wiwa, ati awọn iṣẹ ṣiṣe taagi ni iwọn.
Kọ awọn yiyan
Awọn ẹgbẹ ẹda le ṣe apẹrẹ awọn imọran yiyara pẹlu awọn atunyẹwo afọwọṣe diẹ.
Ẹgbẹ ati ṣiṣan iṣẹ
Awọn iṣẹ ṣiṣe le lo aworan ati awọn ifihan agbara fidio ti o nira tẹlẹ lati ṣiṣẹ.
Ojo iwaju ti Mip-NeRF ati Awọn aaye Radiance Anti-Aliased
Mip-NeRF ṣe ifilọlẹ idile kan ti awọn aaye atako. Mip-NeRF 360 awọn cones ti o gbooro si awọn iwoye ita gbangba ti ko ni opin pẹlu ogun ihamọ, ati Zip-NeRF dapọ anti-aliasing cone-based cone pẹlu awọn aṣoju hash-grid ti o yara lati gba didara mejeeji ati iyara. Reti ero inu-ibanujẹ ti iṣipopada lati tẹsiwaju lilọ kiri si Gaussian splatting ati awọn opo gigun ti akoko gidi, nibiti multiscale, imupadabọ inagijẹ ọfẹ lori awọn foonu ati awọn agbekọri jẹ ibi-afẹde fun AR, maapu, ati gbigba immersive.
Real-World imuse
Ṣiṣe ohun kan ti o mu ni mimọ ni oluwo ọja ti o jẹ ki awọn olumulo sun-un lati iwo yara kikun si isalẹ si alaye dada ti o dara laisi didan.
Atunse awọn iwoye ita gbangba nla (nipasẹ Mip-NeRF 360) fun irin-ajo fojuri ati awọn irin-ajo ohun-ini gidi nibiti kamẹra ti n lọ nipasẹ ọpọlọpọ awọn ijinle.
Ṣiṣẹda aworan ikẹkọ ibaramu ni awọn ipinnu pupọ fun awọn roboti tabi awọn afọwọṣe awakọ adase.
Ṣiṣejade awọn fireemu-iwo aramada sintetiki agaran fun fiimu ati iṣaju VFX nibiti aliasing yoo fọ ibọn naa.
Awọn ewu & Awọn ọna iṣọ
Awọn ẹtọ aworan ati igbanilaaye le di awọn eewu labẹ ofin ti o ba jẹ afihan.
Iṣe awoṣe le yatọ kọja ina, awọn ẹda eniyan, ati awọn agbegbe.
Awọn idaniloju eke le ma ṣe akiyesi ayafi ti a ba ṣe abojuto awọn ala igbẹkẹle.
Ilana Ilana imuse
Ṣetumo awọn ibeere gbigba fun pipe, iranti, ati awọn idiyele aṣiṣe.
Ṣe idanwo pẹlu data ti o baamu awọn ipo iṣelọpọ gidi.
Ṣafikun atunyẹwo eniyan fun igbẹkẹle kekere tabi awọn asọtẹlẹ ipa-giga.
Tọpinpin awoṣe ki o ṣe tunṣe lẹhin kamẹra tabi awọn ayipada datasetto.
Tesiwaju Ṣiṣawari
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 Mip-NeRF and Anti-Aliased Radiance Fields 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
Itọsọna atẹle
Awọn aaye Radiance Neural
Awọn ibeere ti a beere nigbagbogbo
Kini Mip-NeRF ati Anti-Aliased Radiance Fields?
Mip-NeRF ṣe atunṣe blurry, awọn ohun-ọṣọ ti o ni jagidi ti o kọlu NeRF atilẹba nigbati o ba ṣe awọn iwoye ni awọn ọna jijin tabi awọn ipinnu. O ṣe eyi nipa wiwa awọn cones dipo awọn eegun tinrin ailopin, ṣiṣe awọn iṣẹlẹ 3D ti o mu ni didasilẹ ati yiyara lati ṣe ikẹkọ.
Apẹrẹ jiometirika wo ni Mip-NeRF ṣe itopase nipasẹ ibi iṣẹlẹ dipo ray tinrin ti NeRF atilẹba lo?
Mip-NeRF sọ kọnu kan lati inu kamẹra nipasẹ piksẹli kọọkan o si pin si awọn idiwọ conical, yiya iwọn didun ti piksẹli kan ni wiwa gangan.
Kini oruko Mip-NeRF's ifaminsi ti o rọpo fifi koodu ipo boṣewa?
Mip-NeRF ṣe afihan ifidipo ipo isọpọ, eyiti o ṣe koodu agbegbe kan (isunmọ-isunmọ Gaussian) dipo aaye kan.
Kini idi ti NeRF atilẹba ṣe agbejade aliasing nigbati ijinna kamẹra ba yipada?
Apeere aaye tinrin kan kọju ifẹsẹtẹ ti ẹbun kan, nitoribẹẹ awọn piksẹli to sunmọ ati ti o jinna ni a ṣe itọju bakanna, nfa jaggies ati didan.
Bawo ni Mip-NeRF ṣe isunmọ agbegbe ti o wa ninu ibanujẹ conical fun fifi koodu?
Ibanujẹ kọọkan jẹ isunmọ nipasẹ Gaussian pupọ, ati pe fifi koodu ṣe iṣiro iye ti a reti ti awọn sinusoids lori Gaussian yẹn.
Kini yoo ṣẹlẹ si awọn ẹya fifi koodu igbohunsafẹfẹ giga-giga inu frustum nla kan?
Nitoripe awọn sinusoids-igbohunsafẹfẹ yatọ pupọ kọja ibanujẹ nla kan, iṣakojọpọ lori Gaussian nmu wọn lọ si odo, fifun mipmap-bi egboogi-aliasing.