AI Paper Atlas

Research map / 3D vision / 3D generation

Scaling4D: Pushing the Frontier of Video Novel View Synthesis through Large-Scale Monocular Videos

Hongrui Cai, Junjie Luo, Zhihong Fu, Shengnan Zhu, Jiawei Wen, Wanquan Feng, Songtao Zhao, Qian He

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

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 real monocular video data (SpatialVID) to synthetic-only training (ablation)Scaling4D · Panda-70M single-view (400 videos)FVD: 472.62 -> 411.17Improved
add synthetic data to real-only training (ablation)Scaling4D · Panda-70M single-view (400 videos)TransErr: 308.92 -> 286.77Improved
add synthetic data to real-only training (ablation)Scaling4D · Panda-70M single-view (400 videos)FVD: 409.16 -> 411.17No clear effect
add double reprojection (inpainting-style) data (ablation)Scaling4D · Panda-70M single-view (400 videos)RotErr: 6.48 -> 6.27Improved
add double reprojection (inpainting-style) data (ablation)Scaling4D · Panda-70M single-view (400 videos)FVD: 411.17 -> 439.77Hurt
replace point-map inpainting with correspondence-guided generationScaling4D vs TrajectoryCrafter · Panda-70M single-view (400 videos)FID: 68.94 -> 62.83Improved
replace point-map inpainting with correspondence-guided generationScaling4D vs TrajectoryCrafter · iPhone multi-viewLPIPS: 0.539 -> 0.468Improved
increase training data volume from 1k to 3000k videosScaling4D · Panda-70M single-view (400 videos)FID/FVD/CLIP-VImproved
increase training data from 100k to 3000kScaling4D · Panda-70M single-view (400 videos)RotErr/TransErrNo clear effect
train with flow-based correspondence rather than depth-basedScaling4D · 100 static videosEPE: EPE 1.18 px; vec-corr 0.986Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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