Izinkambu ze-Neural Radiance
I-Neural Radiance Fields (NeRF) yakha kabusha isigcawu se-3D esigcwele sisuka ezithombeni ezimbalwa ezijwayelekile, ezikuvumela ukuthi undizise ikhamera iye emibonweni emisha sha.
Uhlolojikelele
It reframed 3D capture as training a tiny neural network rather than building a mesh.
I-Deep Dive
Yethulwe ngo-2020 nguMildenhall nozakwabo, i-NeRF igcina yonke indawo yesehlakalo ngaphakathi kwenethiwekhi encane ye-neural (i-multilayer perceptron). Uma kunikezwe iphoyinti le-3D nendlela yokubuka, inethiwekhi ikhipha umbala walelo phuzu nokuthi i-opaque kangakanani. Ukuze inikeze iphikseli, i-NeRF idubula umsebe endaweni, isampula ikhomba eduze kwayo, ibuze inethiwekhi, futhi ihlanganise imiphumela isebenzisa ukunikezwa kwevolumu. Ngenxa yokuthi yonke le nqubo iyakwazi ukuhlukaniswa, inethiwekhi iqeqeshwa ngokuqhathanisa amaphikseli anikeziwe nezithombe zokufakwayo zangempela futhi ilungiswe ize ifane. Inzuzo i-photorealism emangalisayo, okuhlanganisa imithelela encike ekubukeni efana nokuboniswa nokugqama okucwebezelayo okushintshayo njengoba uhamba. Okubi ukuthi isigcawu ngasinye sidinga ukuqeqeshwa kwaso, futhi indlela yasekuqaleni yayihamba kancane ekuqeqesheni nasekunikezeni.
I-Technical Insight
I-NeRF imele isigcawu njengomsebenzi we-5D oqhubekayo: faka indawo (x, y, z) kanye nendawo yokubuka (ama-engeli amabili), futhi i-MLP ibuyisela umbala we-RGB nokuminyana kwevolumu. Umniningwane obalulekile umbhalo wekhodi wendawo, ohlanganisa amamephu ngokusebenzisa imisebenzi ye-sine ye-high-frequency kanye ne-cosine ukuze inethiwekhi ithwebule imininingwane ecijile esikhundleni sokukhiqiza okukhiphayo okufiphele. Ukunikezwa kuhlanganisa umbala nokuminyana emisebeni yekhamera ngayinye, ukulinganisa eduze, amasampuli afiphele kakhulu kakhulu, ngqo izibalo zokunikezwa kwevolumu yakudala ezenziwe zaqeqesheka.
I-Strategic Impact
Isivinini nesikali
I-Visual AI ingakwazi ukuhlola, ukutholwa, nokumaka imisebenzi esikalini.
Yakha ukukhetha
Amathimba aqanjiwe angakwazi ukulinganisa imiqondo ngokushesha ngezibuyekezo ezimbalwa ezenziwa mathupha.
Ithimba kanye nokusebenza komsebenzi
Imisebenzi ingasebenzisa amasiginali wesithombe nawevidiyo obekunzima ukuwenza ngaphambilini.
Ikusasa Lezinkambu Ze-Neural Radiance
Ucwaningo lwe-NeRF luqhume ngemva kuka-2020, nokulandelwa okufana nokuqeqeshwa kokusika i-Instant-NGP kusuka emahoreni kuya kumasekhondi kusetshenziswa amakhodi wegridi ye-hash, kanye ne-Mip-NeRF ithuthukisa ikhwalithi kuzo zonke izikali. Inkambu iya ngokuya ihlanganiswa noma iphonselwa inselelo yi-Gaussian Splatting, enikezela ngokushesha. Lindela amasu asuselwe ku-NeRF ekwenzeni imephu, ukubukwa komkhiqizo we-e-commerce, imiphumela yokubuka yefilimu, ne-AR/VR, kanye nokukhula kwama-NeRF ashukumisayo aphatha izigcawu ezinyakazayo kanye nokuthwebula "endle" ngokushintsha ukukhanya. Amatimu amakhulu isivinini, ukuhleleka, nokuthwebula izigcawu ezivela ezithombeni ezimbalwa, ezilula.
Ukuqaliswa Komhlaba Wangempela
Ukuguqula ividiyo yefoni yento ibe umbono we-3D ongayizungeza ukuze uthenge ku-inthanethi
Ukwakha kabusha izindawo zangempela njengengemuva le-photorealistic lefilimu nemiphumela ebonakalayo
Ukwakha izigcawu ezigxilile ze-3D zokuzizwisa okungokoqobo nokungathandwa kwabathelisi esikubona
Igcina ngokudijithali amasayithi amagugu amasiko nezinto zobuciko ezivela kumasethi ezithombe
Izingozi & Guardrails
Amalungelo ezithombe kanye nemvume kungaba ubungozi bezomthetho uma ukuvela kungacacile.
Ukusebenza kwemodeli kungahluka kukho konke ukukhanya, izibalo zabantu, kanye nezindawo.
Okuhle okungelona iqiniso kungase kungabonakali ngaphandle uma izinga lokuzethemba liqashelwa.
Ukuqalisa Umhlahlandlela
Chaza indlela yokwamukela yokunemba, ukukhumbula, nezindleko zamaphutha.
Hlola ngedatha efana nezimo zangempela zokukhiqiza.
Engeza isibuyekezo somuntu ukuze uthole ukuzethemba okuphansi noma izibikezelo zomthelela omkhulu.
Landelela ukukhukhuleka kwemodeli bese uqinisekisa kabusha ngemva kwezinguquko zekhamera noma zesethi yedatha.
Qhubeka Uhlole
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 Neural 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
Umhlahlandlela olandelayo
I-Mip-NeRF kanye ne-Anti-Aliased Radiance Fields
Imibuzo evame ukubuzwa
What is Neural Radiance Fields?
I-Neural Radiance Fields (NeRF) yakha kabusha isigcawu se-3D esigcwele sisuka ezithombeni ezimbalwa ezijwayelekile, ezikuvumela ukuthi undizise ikhamera iye emibonweni emisha sha. Ifake kabusha ukuthwebula kwe-3D njengokuqeqesha inethiwekhi encane ye-neural kunokwakha i-mesh.
Uyini umgomo oyinhloko we-Neural Radiance Field?
I-NeRF yakha ukumelwa kwe-3D kusuka kusethi yezithombe, okuvumela ukunikezwa kwe-photorealistic kusuka emibonweni engekho ezithombeni zangempela.
Ithini i-NeRF neural network output okokufaka okunikeziwe?
Ukuze uthole indawo ye-3D nesiqondiso sokubuka, i-MLP ibuyisela umbala nokuminyana (i-opacity) ngaleso sikhathi.
Yiziphi izinto ezifakwayo kumsebenzi we-5D we-NeRF?
I-NeRF ithatha indawo ye-3D (x, y, z) kanye nendawo yokubuka (ama-engeli amabili), okuhlanganisa okokufaka okuhlanu.
I-NeRF iyiguqula kanjani imiphumela yayo yenethiwekhi ibe umbala wephikseli wokugcina?
Amasampula e-NeRF akhomba eduze kwereyi ngalinye futhi ahlanganisa umbala wawo olinganiselwe ngokuminyana, isibalo sakudala sokunikeza ivolumu eyenziwe yahlukaniswa.
Kungani i-NeRF isebenzisa umbhalo wekhodi wendawo kuzixhumanisi zokufaka?
Izixhumanisi zemephu ngokusebenzisa imisebenzi ye-sine/cosine ye-high-frequency ivumela i-MLP imele imininingwane emihle esikhundleni semiphumela engacacile.