InoteveraGaidhi rinotevera
Kukurukurirana Mufananidzo Kugadziriswa neMultimodal Models
AI Yekutarisa
Gwara reVisual AI
AI image relighting changes the lighting of an existing photo, for example moving the light source, changing its color or matching a new background, while keeping the subject's identity and shape.
Teerera kugwaro iriAI-yakarondedzerwa · ~3 min
Tools such as IC-Light use diffusion models to re-render shading and highlights. This matters for product photography, portrait editing and compositing, where mismatched light is one of the most obvious signs of a fake or pasted-in image.
What you see in a photo is roughly the result of what a surface is (its base color, or albedo, its shape and its material) combined with the light falling on it. Relighting means changing the light while leaving everything else alone. That is hard because a single photo mixes these factors together. Classical and early learned methods tried to separate the factors explicitly. Intrinsic image decomposition splits an image into albedo and shading. Other models estimate surface normals and depth, then re-shade the scene with a new light described by an environment map or low-order spherical harmonics. Portrait relighting research, such as Google's Total Relighting (2021), trained on data from light stages, which are domes of controllable lights that record a person under many lighting conditions. These methods can be precise but struggle outside their training data, for example with hair, glass or complex scenes. Diffusion-based relighting takes a different approach: a generative model re-renders the image under the requested lighting. IC-Light (Imposing Consistent Light) was released in 2024 by Lvmin Zhang, the developer behind ControlNet. It comes in versions conditioned on a text description of the light or on a new background image. Its key training idea is consistent light transport. In physics, an object's appearance under a mix of two lights equals the mix of its appearances under each light alone. Enforcing this during training helps the model change the lighting without changing the object's identity or details. Two misconceptions are common. First, relighting is not a color filter: a filter cannot move shadows or add highlights on the side facing a new light. Second, the model does not recover true 3D geometry. It produces plausible shading, which can fail on text, fine patterns and reflective surfaces, or when the requested lighting is physically inconsistent. Results can also subtly change details such as skin texture, so inspect outputs closely.
An online shop photographs products under plain studio light, then relights each one to match lifestyle backgrounds so shadows and highlights look consistent.
A portrait editor adds warm window light from the left to a flatly lit headshot, using a text prompt such as 'sunlight from left window' with IC-Light.
A compositor cuts a person out of one photo and places them on a neon-lit street. They use a background-conditioned relighting model so the person picks up the scene's colored light.
Mufirimu anogadzira storyboard anoedza mamiriro ezuva akasiyana, kubva panguva ye golden hour kusvika pakuvharwa kwemakore, pane imwecheteyo foto yenzvimbo asati atanga kupfura.
Relighting inowedzera kubva pamifananidzo imwe chete kusvika kumavhidhiyo ne3D, uko chiedza chinofanira kuramba chakafanana mumifananidzo nemaziso akasiyana. Tsvakurudzo inosanganisa maitiro ekugadzira aine ongororo dzakajeka dzezviyero nematani kupa vashandisi kudzora kwepanyama, sekuisira chiedza panzvimbo yakatarwa. Relighting inobatawo maonero echokwadi. Mavara asina kusimba akanga ari chiratidzo chekuona zvigadzirwa, saka relighting iri nani inowedzera kushaikwa kwekuziva, izvo zvinowedzera kusimba kwechikumbiro chechigadzirwa chakakodzera. Kune vashandisi vezvinhu, shanduko inowanikwa munguva pfupi inowanikwa ndeyekuti maitiro e relighting achava akajairika mumagadzirirwo emazuva ano.
Tsamba yemahara
Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.
Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.
Zviedze iwe pachako
Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.
Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo
AI image relighting inochinja chiedza chemufananidzo uripo, semuenzaniso kufambisa chiedza, kuchinja ruvara rwayo kana kuenzanisa nekunyorerwa kutsva, ichichengetedza kuzivikanwa kwechinhu uye chimiro. Zvishandiso zvakaita se IC-Light anoshandisa maitiro ekudhindisa kudzokorora kusvibisa uye kusvibisa. Izvi zvakakosha pakufotora zvigadzirwa, kugadzirisa mifananidzo uye kusanganisa, uko chiedza chisina kusangana ndicho chimwe chezviratidzo zvakajeka zvechokwadi kana mufananidzo wakaiswa.
Relighting inochinja chete chiedza ichichengetedza izvo zviri pamusoro, kusanganisira ruvara rwayo rwepamusoro, chimiro uye zvinhu.
Inopatsanura mufananidzo kuita albedo, ruvara rwepamusoro rwepamusoro, uye shading, maitiro e chiedza pane iyo surface.
Matanho ezuva imadome ane mwenje unokwanisa kudzora anobata maitiro emunhu ari pasi pemwenje yakasiyana-siyana, achipa data rekudzidzisa nzira dzakadai seTotal Relighting.
Kusarongeka uku kwekutakurwa kwemwenje kunobatsira modhi kuchinja mwenje pasina kushandura hunhu hwechinhu kana zvimwe zvakadzama.
Chidzitiro chinogadzirisa ruvara rese mufananidzo wese. Kudzokorodza kwechokwadi kunochinja nzvimbo dzemumvuri nemwenje zvichienderana nenzira yemwenje.
Ramba uchidzidza
Mamwe madhairekitori akasarudzirwa nyaya iyi
InoteveraGaidhi rinotevera
Kukurukurirana Mufananidzo Kugadziriswa neMultimodal Models
AI Yekutarisa