---
title: "ViSA-Gait: Leveraging Vision Foundation Models for Semantic Anchored Gait Recognition"
authors:
  - "Xiangru Li"
  - "Deqiang Yin"
  - "Yifan Xie"
  - "Guojian Li"
  - "Zebang Cheng"
  - "Fei Ma"
doi: "10.24963/ijcai.2026/151"
url: "https://ijcai.org/proceedings/2026/151"
pdf: "https://ijcai.org/proceedings/2026/0151.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "1351-1359"
track: "Main Track"
keywords:
  - "Computer Vision: Recognition (object detection, categorization)"
  - "Computer Vision: Representation learning"
  - "Computer Vision: Low-level Vision"
  - "Computer Vision: Machine learning for vision"
---

# ViSA-Gait: Leveraging Vision Foundation Models for Semantic Anchored Gait Recognition

Xiangru Li, Deqiang Yin, Yifan Xie, Guojian Li, Zebang Cheng, Fei Ma

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 1351-1359. DOI: [10.24963/ijcai.2026/151](https://doi.org/10.24963/ijcai.2026/151)

## Abstract

Gait recognition has achieved remarkable success in constrained environments, yet its performance often degrades significantly in cross-domain and cross-vertical-view scenarios. This is primarily due to the fact that domain-specific silhouette geometry causes models to overfit to extrinsic geometric characteristics rather than learning generalizable motion patterns. To address this issue, we present ViSA-Gait, which utilizes Vision Foundation Models (VFMs) as a “Semantic Compass” for stable universal human body knowledge guidance, enabling a lightweight backbone to robustly capture fine-grained gait dynamics. Specifically, we introduce a token-based distillation mechanism where a Spatial Token Learner (STL) and a Temporal Token Learner (TTL) filter dense VFM features into motion-consistent descriptors. These descriptors are then adaptively injected into a lightweight 3D-CNN backbone via a Gated Cross-Attention (GCA) mechanism, functioning as “Semantic Anchors” that regularize the feature space and guide the model to focus on intrinsic body motion. Extensive experiments demonstrate that ViSA-Gait achieves SOTA performance on cross-domain and cross-vertical-view benchmarks while remaining competitive within-domain, offering a new perspective on bridging the gap between geometry-based analysis and universal semantic understanding.