GeoFix: Geometry Fixing for 4D Consistent Generation

Hugging Face Daily Papers Papers
4d-generation video-diffusion geometric-consistency reward-alignment depth-estimation

Summary

GeoFix提出了一种统一框架，通过单一几何奖励提升4D生成中的几何一致性和视觉质量，该奖励衡量静态区域的跨视角深度一致性，并针对视角重叠效应进行校准，实现了WorldScore 3D一致性提升39%和VBench质量提升4%。

Video diffusion models increasingly serve as priors for 4D generation, but the generated videos often lack geometric consistency. Recent approaches address this through post-training by designing geometric rewards, yet this can degrade texture and fine detail, requiring auxiliary appearance rewards and careful reward balancing. To address this challenge, we propose GeoFix, a unified framework that improves geometric consistency and visual quality with a single geometric reward, applicable to both training and inference-time reward alignment. Our reward measures cross-view depth consistency in static scene regions, targeting structural inconsistencies without directly constraining appearance or penalizing valid object motion. Crucially, depth agreement increases with camera field-of-view overlap even in real videos and 4D datasets, allowing a smaller camera movement to be mistaken for better geometry. We account for this effect by calibrating the reward against the overlap-dependent trend. GeoFix achieves a 39% improvement in WorldScore 3D consistency and a 4% increase in VBench Quality Score.
Original Article
View Cached Full Text

Cached at: 10/14/26, 06:53 AM

Paper page - GeoFix: Geometry Fixing for 4D Consistent Generation

Source: https://huggingface.co/papers/2610.14293

Abstract

Videodiffusionmodelsincreasinglyserveaspriorsfor4Dgeneration,butthegeneratedvideosoftenlackgeometricconsistency.Recentapproachesaddressthisthroughpost-trainingbydesigninggeometricrewards,yetthiscandegradetextureandfinedetail,requiringauxiliaryappearancerewardsandcarefulrewardbalancing.Toaddressthischallenge,weproposeGeoFix,aunifiedframeworkthatimprovesgeometricconsistencyandvisualqualitywithasinglegeometricreward,applicabletobothtrainingandinference-timerewardalignment.Ourrewardmeasurescross-viewdepthconsistencyinstaticsceneregions,targetingstructuralinconsistencieswithoutdirectlyconstrainingappearanceorpenalizingvalidobjectmotion.Crucially,depthagreementincreaseswithcamerafield-of-viewoverlapeveninrealvideosand4Ddatasets,allowingasmallercameramovementtobemistakenforbettergeometry.Weaccountforthiseffectbycalibratingtherewardagainsttheoverlap-dependenttrend.GeoFixachievesa39%improvementinWorldScore3Dconsistencyanda4%increaseinVBenchQualityScore.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2610\.14293

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.14293 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.14293 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.14293 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space

Hugging Face Daily Papers

GeoVerse is a framework for synthesizing world-consistent novel views by performing generation in a geometric latent space from a 3D foundation model and injecting appearance priors from a video generative model, demonstrating improved visual quality and geometric consistency over existing methods.

WorldRecon: Generative Multi-View Mesh Reconstruction for Large-Scale 3D World

Hugging Face Daily Papers

WorldRecon 提出一种生成式多视角网格重建框架，可从无位姿 RGB 图像重建大规模 3D 场景的完整网格。它将 TRELLIS.2 生成模型与重建式条件架构结合，通过邻域感知潜在修补和因果世界生成保证场景级一致性，在合成与真实室内场景上分别将 Chamfer 距离降低 14.9% 和 43.1%，达到最先进性能。

Quantitative Video World Model Evaluation for Geometric-Consistency

Hugging Face Daily Papers

A quantitative framework called PDI-Bench is introduced for evaluating geometric coherence in generated videos through monocular reconstruction and projective-geometry residuals, revealing geometry-specific failure modes in video generators.