DAGait: Generalized Skeleton-Guided Data Alignment for Gait Recognition

2025-03-24

AI Summary

Existing gait recognition methods suffer significant performance degradation in unconstrained, real-world scenarios, primarily due to spatiotemporal distribution shifts—such as variations in viewpoint, subject position, and distance—that induce silhouette deformation. To address this, we propose a skeleton-prior-guided affine alignment strategy: leveraging human skeletal keypoints to establish geometric constraints that drive affine transformations for normalizing silhouette scale, rotation, and translation, thereby enhancing cross-domain robustness. This work is the first to systematically reveal the critical impact of data alignment on the generalization capability of gait recognition. We further introduce a generic, plug-and-play alignment module compatible with diverse network architectures. Evaluated on the Gait3D dataset, our method achieves an average accuracy improvement of 7.9% and up to 24.0% gain in cross-domain recognition, demonstrating both effectiveness and strong generalizability.

Technology Category

• Computer Vision: Motion & Tracking
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Gait recognition is emerging as a promising and innovative area within the field of computer vision, widely applied to remote person identification. Although existing gait recognition methods have achieved substantial success in controlled laboratory datasets, their performance often declines significantly when transitioning to wild datasets.We argue that the performance gap can be primarily attributed to the spatio-temporal distribution inconsistencies present in wild datasets, where subjects appear at varying angles, positions, and distances across the frames. To achieve accurate gait recognition in the wild, we propose a skeleton-guided silhouette alignment strategy, which uses prior knowledge of the skeletons to perform affine transformations on the corresponding silhouettes.To the best of our knowledge, this is the first study to explore the impact of data alignment on gait recognition. We conducted extensive experiments across multiple datasets and network architectures, and the results demonstrate the significant advantages of our proposed alignment strategy.Specifically, on the challenging Gait3D dataset, our method achieved an average performance improvement of 7.9% across all evaluated networks. Furthermore, our method achieves substantial improvements on cross-domain datasets, with accuracy improvements of up to 24.0%.

Problem

• Addresses performance decline in gait recognition from lab to wild datasets
• Proposes skeleton-guided silhouette alignment for spatio-temporal inconsistencies
• Improves cross-domain accuracy by up to 24.0% in gait recognition

Innovation

• Skeleton-guided silhouette alignment strategy
• Affine transformations using skeleton knowledge
• Improves cross-domain accuracy significantly