AI Paper Atlas

Research map / 3D vision / 3D generation

SpatialCrafter: Unleashing the Imagination of Video Diffusion Models for Scene Reconstruction from Limited Observations

Songchun Zhang, Huiyao Xu, Sitong Guo, Zhongwei Xie, Hujun Bao, Weiwei Xu, Changqing Zou

ICCV 2025 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 10 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 ray embedding camera encoding (ablation)SpatialCrafter video diffusion · Video diffusion ablationFVD: 112.843 -> 78.132Improved
use ray embedding instead of raw value embedding (ablation)SpatialCrafter video diffusion · Video diffusion ablationT_err: 1.592 -> 1.014Improved
add epipolar attention (ablation)SpatialCrafter video diffusion · Video diffusion ablationCOLMAP_err: 12.2% -> 4.2%Improved
add epipolar attention (ablation)SpatialCrafter video diffusion · Video diffusion ablationFVD: 79.547 -> 78.132Improved
add unified scale alignment across datasets (ablation)SpatialCrafter video diffusion · Video diffusion ablationT_err: 1.712 -> 1.014Improved
use SpatialCrafter instead of ViewCrafter for generationcamera-controlled video generation · Tanks-and-TemplesPSNR: 16.24 -> 16.12Mixed
use SpatialCrafter instead of ReconX for two-view reconstructionRealEstate10K two-view · RealEstate10KLPIPS: 0.088 -> 0.121Hurt
use SpatialCrafter instead of MVSplat on low-overlap inputsRealEstate10K two-view small overlap · RealEstate10K (small overlap)PSNR: 20.353 -> 22.514Improved
use SpatialCrafter instead of CAT3D on DTU 3-viewsparse-view reconstruction · DTU (3-view)PSNR: 22.02 -> 27.92Improved
add video-latent generative priors and depth features to reconstruction (ablation)feed-forward Gaussian reconstruction · qualitativegeometric detail fidelityImproved

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