Skip to content

IllumiNeRF

3D relighting without inverse rendering

Visit official site

What IllumiNeRF is

The technique is described as a way to relight 3D scenes by first relighting each input image with an image diffusion model, then rebuilding a Neural Radiance Field for target lighting.

What you can do with it

  • Recover a 3D representation from images
  • Avoid the cost of inverse rendering
  • Render new views under target illumination

Getting started

  • Gather images under different lighting
  • Relight the inputs with the diffusion model
  • Rebuild the NeRF and set the target lighting

Pros & cons

✓ What we like

  • Avoids costly differentiable Monte Carlo rendering
  • Renders new views under chosen lighting

! What to watch out for

  • Input is a set of images under varied lighting
  • Results are reported on relighting benchmarks

FAQ

What does it avoid?

The material says it avoids inverse rendering with differentiable Monte Carlo optimisation.

How are new views made?

The steps say to rebuild a NeRF and set the target lighting.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review