Measuring 3D Spatial Geometric Consistency in Dynamic Generated Videos

TL;DR

This paper introduces SGC, a new metric for evaluating 3D spatial geometric consistency in dynamic videos, addressing limitations of existing metrics by accurately detecting geometric inconsistencies.

Contribution

We propose a novel metric, SGC, that measures divergence among local camera poses to evaluate 3D geometric consistency in generated videos.

Findings

SGC effectively detects geometric inconsistencies in generated videos.

SGC outperforms existing metrics in identifying critical geometric failures.

Experiments on real and synthetic videos validate SGC's robustness.

Abstract

Recent generative models can produce high-fidelity videos, yet they often exhibit 3D spatial geometric inconsistencies. Existing evaluation methods fail to accurately characterize these inconsistencies: fidelity-centric metrics like FVD are insensitive to geometric distortions, while consistency-focused benchmarks often penalize valid foreground dynamics. To address this gap, we introduce SGC, a metric for evaluating 3D \textbf{S}patial \textbf{G}eometric \textbf{C}onsistency in dynamically generated videos. We quantify geometric consistency by measuring the divergence among multiple camera poses estimated from distinct local regions. Our approach first separates static from dynamic regions, then partitions the static background into spatially coherent sub-regions. We predict depth for each pixel, estimate a local camera pose for each subregion, and compute the divergence among these poses to quantify geometric consistency. Experiments on real and generative videos demonstrate that SGC robustly quantifies geometric inconsistencies, effectively identifying critical failures missed by existing metrics.