AI Paper Atlas

Research map / 3D vision / 3D generation

From Rays to Projections: Better Inputs for Feed-Forward View Synthesis

Zirui Wu, Zeren Jiang, Martin R. Oswald, Jie Song

CVPR 2026 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 5 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 projective conditioning (point cloud projection image) to the LVSM baseline (ablation)LVSM 12 layers · RealEstate10K (Total)PSNR: 24.60 -> 25.20Improved
add DINOv3 feature prior on top of projection (ablation)PVSM without pretraining · RealEstate10K (Total)PSNR / LPIPS: PSNR 25.20 -> 25.13, LPIPS 0.177 -> 0.163Mixed
add MAE pretraining on DL3DV (ablation)PVSM with DINO and projection · RealEstate10K (Total)PSNR: 25.13 -> 25.64Improved
pretrain on large DL3DV instead of RealEstate10K (ablation)PVSM 12 layers · RealEstate10K (Total)PSNR: 24.78 -> 25.64Improved
increase finetuning from 200 steps to 50K steps after DL3DV pretraining (ablation)PVSM 12 layers · RealEstate10K (Total)PSNR: 22.32 -> 25.64Improved
add RoPE to the projective model (ablation)PVSM · Single-object overfitPSNR: 21.18 -> 30.03Improved
add RoPE to LVSM (ablation)LVSM · Single-object overfitPSNR: 25.39 -> 25.88Improved
replace Plucker ray conditioning with projective conditioningLVSM · Consistency Benchmark (World Scale)PSNR (M): 14.56 -> 25.43Improved
fine-tune with 500 steps of camera augmentationLVSM · Consistency Benchmark (World Scale)PSNR (M): 14.56 -> 13.57Hurt
fine-tune with 500 steps of camera augmentationPVSM · Consistency Benchmark (World Scale)PSNR (M): 25.43 -> 25.78Improved
projective conditioning vs Plucker rays under rollPVSM vs LVSM · Consistency Benchmark (Roll)PSNR (M): 19.54 -> 17.53Hurt
scale model from 12 to 24 layersPVSM · RealEstate10K (Total)PSNR: 25.64 -> 26.90Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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