Walking Further: Semantic-aware Multimodal Gait Recognition Under Long-Range Conditions

2026-03-14

AI Summary

Existing gait recognition methods are largely confined to short-range, single-modality scenarios and struggle to address the challenges of long-range and cross-range recognition in real-world settings. To bridge this gap, this work introduces LRGait, the first LiDAR-camera multimodal benchmark specifically designed for long-range gait recognition, along with an end-to-end framework named EMGaitNet. EMGaitNet effectively mitigates the modality gap between 2D visual and 3D geometric features through CLIP-driven semantic mining, semantic-guided cross-modal alignment, and a symmetric cross-attention fusion mechanism, while incorporating spatiotemporal dynamic modeling to enhance robustness. Extensive experiments demonstrate that EMGaitNet significantly outperforms state-of-the-art methods across multiple datasets, achieving particularly superior performance in long-range and cross-range scenarios.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Gait recognition is an emerging biometric technology that enables non-intrusive and hard-to-spoof human identification. However, most existing methods are confined to short-range, unimodal settings and fail to generalize to long-range and cross-distance scenarios under real-world conditions. To address this gap, we present \textbf{LRGait}, the first LiDAR-Camera multimodal benchmark designed for robust long-range gait recognition across diverse outdoor distances and environments. We further propose \textbf{EMGaitNet}, an end-to-end framework tailored for long-range multimodal gait recognition. To bridge the modality gap between RGB images and point clouds, we introduce a semantic-guided fusion pipeline. A CLIP-based Semantic Mining (SeMi) module first extracts human body-part-aware semantic cues, which are then employed to align 2D and 3D features via a Semantic-Guided Alignment (SGA) module within a unified embedding space. A Symmetric Cross-Attention Fusion (SCAF) module hierarchically integrates visual contours and 3D geometric features, and a Spatio-Temporal (ST) module captures global gait dynamics. Extensive experiments on various gait datasets validate the effectiveness of our method.

Problem

• gait recognition
• long-range
• multimodal
• cross-distance
• biometric identification

Innovation

• multimodal gait recognition
• long-range perception
• semantic-guided fusion
• LiDAR-camera fusion
• CLIP-based semantic mining