AI Paper Atlas

Research map / 3D vision / Gaussian splatting

YoNoSplat: You Only Need One Model for Feedforward 3D Gaussian Splatting

Botao Ye, Boqi Chen, Haofei Xu, Daniel Barath, Marc Pollefeys

ICLR 2026 · 3D vision › Gaussian splatting · approach: Compact and efficient Gaussian splatting · 27 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
use mix-forcing instead of self-forcing for pose aggregation (ablation)YoNoSplat · 6-view ablation, pose-freePSNR: 24.652 -> 25.587Improved
use mix-forcing instead of teacher-forcing (ablation)YoNoSplat · 6-view ablationPSNR: pose-free 25.300 -> 25.587; pose-dependent 25.228 -> 25.212Mixed
normalize scene by max pairwise camera distance instead of none (ablation)YoNoSplat · 6-view ablationPSNR: 22.662 -> 25.212Improved
max pairwise distance vs mean pairwise distance normalization (ablation)YoNoSplat · 6-view ablationPSNR: 24.950 -> 25.212Improved
predict local Gaussians instead of canonical-space Gaussians (ablation)YoNoSplat · 6-view ablationPSNR: 24.104 -> 25.587Improved
add intrinsic conditioning (predicted intrinsics) via ICE (ablation)YoNoSplat · 6-view ablationPSNR: 24.481 -> 24.711Improved
use ground-truth intrinsics instead of predicted (ablation)YoNoSplat · 6-view ablationPSNR: 24.711 -> 25.587Improved
add Plucker ray embedding (ablation)YoNoSplat · 6-view ablationPSNR: 25.212 -> 25.202No clear effect
add optional post-optimization of poses and GaussiansYoNoSplat · DL3DV 6 viewsPSNR: 24.531 -> 27.533 (vs InstantSplat 21.677)Improved
supply ground-truth poses at inferenceYoNoSplat · DL3DV 6/24 viewsPSNR: 6v: 24.887 (pose-free) vs 24.717 (GT pose); 24v: 22.354 vs 22.664Mixed
increase number of input viewsYoNoSplat on ScanNet++ · ScanNet++ zero-shotPSNR: 17.935 -> 19.284 (32v to 128v, GT k)Improved
increase input resolution to 518x280 (ablation)YoNoSplat pose estimation · RealEstate10KAUC@5deg: 0.722 -> 0.813Improved

Later papers that compared with this method

2 later accepted papers report YonoSplat as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

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