AI Paper Atlas

Research map / 3D vision / Gaussian splatting

SplatPainter: Interactive Authoring of 3D Gaussians from 2D Edits via Test-Time Training

Yang Zheng, Hao Tan, Kai Zhang, Peng Wang, Leonidas Guibas, Gordon Wetzstein, Wang Yifan

ECCV 2026 · 3D vision › Gaussian splatting · approach: Regularization and primitive design for 3DGS · 6 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
add stage II TTT refinement on top of stage I (ablation)SplatPainter · Local refinement (TexVerse, 370 assets)PSNR: 20.84 -> 21.07Improved
add stage II TTT refinement on top of stage I (ablation)SplatPainter · Local refinement (TexVerse, 370 assets)LPIPS: 0.3782 -> 0.3673Improved
add TTT refinement instead of cross-attention (ablation)SplatPainter · Global relightingPSNR: 24.05 -> 24.56Improved
provide more relit views (1 to 4)SplatPainter · Global relightingPSNR: 24.56 -> 25.03Improved
provide 2 relit views instead of 1SplatPainter · Global relightingLPIPS: 0.0721 -> 0.0719No clear effect
use LocalAtt + feats query instead of VoxelAtt (ablation)Voxel compression · Voxel compression (stage I)PSNR: 30.65 -> 34.02Improved
use feats query instead of latent query (ablation)Voxel compression · Voxel compression (stage I)PSNR: 33.84 -> 34.02Improved
voxel compression versus uncompressed GS-LRMSplatPainter · Voxel compression (stage I)PSNR: 34.19 -> 34.02Hurt
use SplatPainter local TTT update instead of VoxHammerCreative editing · Creative editing, non-edited regionPSNR: 13.86 -> 33.02Improved
use SplatPainter instead of 3DGS optimizationLocal refinement · Local refinement (TexVerse, 370 assets)time (s): 323 -> 1.54+0.64Improved
use SplatPainter instead of SRGSLocal refinement · Local refinement (TexVerse, 370 assets)SSIM: 0.6759 -> 0.5780Hurt

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