Causality-inspired Discriminative Feature Learning in Triple Domains for Gait Recognition

TL;DR

This paper introduces CLTD, a causality-inspired feature learning framework for gait recognition that effectively disentangles identity features from confounders across spatial, temporal, and spectral domains, improving accuracy.

Contribution

The paper proposes a novel causality-inspired module with attention and spectral projection to enhance gait recognition by eliminating confounders and enforcing semantic consistency.

Findings

Significant performance improvements on challenging datasets.

Effective disentanglement of identity features from confounders.

Seamless integration into existing gait recognition methods.

Abstract

Gait recognition is a biometric technology that distinguishes individuals by their walking patterns. However, previous methods face challenges when accurately extracting identity features because they often become entangled with non-identity clues. To address this challenge, we propose CLTD, a causality-inspired discriminative feature learning module designed to effectively eliminate the influence of confounders in triple domains, \ie, spatial, temporal, and spectral. Specifically, we utilize the Cross Pixel-wise Attention Generator (CPAG) to generate attention distributions for factual and counterfactual features in spatial and temporal domains. Then, we introduce the Fourier Projection Head (FPH) to project spatial features into the spectral space, which preserves essential information while reducing computational costs. Additionally, we employ an optimization method with contrastive learning to enforce semantic consistency constraints across sequences from the same subject. Our approach has demonstrated significant performance improvements on challenging datasets, proving its effectiveness. Moreover, it can be seamlessly integrated into existing gait recognition methods.