GeCo: Evaluating Geometric Consistency for Video Generation via Motion and Structure

TL;DR

GeCo is a geometry-based metric that detects geometric and occlusion artifacts in static scenes, aiding benchmarking and improving video generation quality.

Contribution

We introduce GeCo, a novel geometry-grounded metric that jointly detects artifacts and guides video generation without additional training.

Findings

GeCo produces interpretable, dense consistency maps revealing artifacts.

Using GeCo as guidance reduces deformation artifacts in video generation.

Benchmarking with GeCo uncovers common failure modes in recent models.

Abstract

We introduce GeCo, a geometry-grounded metric for jointly detecting geometric deformation and occlusion-inconsistency artifacts in static scenes. By fusing residual motion and depth priors, GeCo produces interpretable, dense consistency maps that reveal these artifacts. We use GeCo to systematically benchmark recent video generation models, uncovering common failure modes, and further employ it as a training-free guidance loss to reduce deformation artifacts during video generation.