AI Paper Atlas

Research map / 3D vision / 3D generation

GeoRelight: Learning Joint Geometrical Relighting and Reconstruction with Flexible Multi-Modal Diffusion Transformers

Yuxuan Xue, Ruofan Liang, Egor Zakharov, Timur Bagautdinov, Chen Cao, Giljoo Nam, Shunsuke Saito, Gerard Pons-Moll, Javier Romero

CVPR 2026 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
jointly generate geometry with the relit image (vs relit image only) (ablation)GeoRelight · synthetic relightingPSNR: 21.19 -> 27.49Improved
jointly generate relit image with geometry versus giving ground-truth geometry (ablation)GeoRelight · synthetic relightingPSNR: 26.96 (GT geometry) vs 27.49 (joint)No clear effect
jointly generate normal and geometry with the relit appearance (vs without appearance) (ablation)GeoRelight · synthetic normal estimationangular error: 12.24 -> 9.10Improved
jointly generate point cloud with relit appearance (vs without appearance) (ablation)GeoRelight · synthetic point cloudChamfer distance: 1.00 -> 0.58Improved
give ground-truth relit image as clear condition for geometry (ablation)GeoRelight · synthetic normal estimationangular error: 9.10 -> 8.55Improved
add in-the-wild data to Synth+Dome mixed training (ablation)GeoRelight · in-the-wild relighting (qualitative)lighting balanceImproved
add Dome (light stage) data to synthetic-only training (ablation)GeoRelight · in-the-wild relighting (qualitative)realismMixed
use joint multi-modal GeoRelight instead of DiffusionRendererrelighting · synthetic dataPSNR: 19.28 -> 27.22Improved
GeoRelight versus NeuralGaffer on HumanOLATrelighting · HumanOLATSSIM: 0.948 (NeuralGaffer) vs 0.935Mixed
use GeoRelight iNOD joint geometry instead of VGGT/MoGe2geometry estimation · synthetic dataChamfer distance: 3.37 -> 0.766Improved
use GeoRelight instead of DiffusionRenderer for albedoalbedo estimation · synthetic dataPSNR: 19.46 -> 28.07Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy