Robust 4D Visual Geometry Transformer with Uncertainty-Aware Priors

TL;DR

This paper introduces a novel framework for dynamic 4D scene reconstruction that effectively disentangles motion and static components using uncertainty-aware priors and attention mechanisms, outperforming existing methods.

Contribution

The proposed approach combines entropy-guided subspace projection, local geometry purification, and uncertainty-aware cross-view consistency to improve dynamic scene reconstruction without scene-specific tuning.

Findings

Reduces Mean Accuracy error by 13.43% on benchmarks.

Improves segmentation F-measure by 10.49%.

Outperforms state-of-the-art methods in dynamic 4D reconstruction.

Abstract

Reconstructing dynamic 4D scenes is an important yet challenging task. While 3D foundation models like VGGT excel in static settings, they often struggle with dynamic sequences where motion causes significant geometric ambiguity. To address this, we present a framework designed to disentangle dynamic and static components by modeling uncertainty across different stages of the reconstruction process. Our approach introduces three synergistic mechanisms: (1) Entropy-Guided Subspace Projection, which leverages information-theoretic weighting to adaptively aggregate multi-head attention distributions, effectively isolating dynamic motion cues from semantic noise; (2) Local-Consistency Driven Geometry Purification, which enforces spatial continuity via radius-based neighborhood constraints to eliminate structural outliers; and (3) Uncertainty-Aware Cross-View Consistency, which formulates multi-view projection refinement as a heteroscedastic maximum likelihood estimation problem, utilizing depth confidence as a probabilistic weight. Experiments on dynamic benchmarks show that our approach outperforms current state-of-the-art methods, reducing Mean Accuracy error by 13.43\% and improving segmentation F-measure by 10.49\%. Our framework maintains the efficiency of feed-forward inference and requires no task-specific fine-tuning or per-scene optimization.