WiGR-AF: Attention Fusion Network for Gait Identity Recognition Using WiFi

With the popularization and development of Wi-Fi devices, identity recognition based on wireless signals has become an important research topic in the fields of ubiquitous computing and human-computer interaction. However, current gait identity recognition based on WiFi Channel State Information (CSI) mostly relies on manually extracted features and requires subjects to walk on a fixed path, which limits its application and promotion in smart home scenarios. To address this issue, we propose a gait recognition method called WiGR-AF based on wireless signals and attention fusion networks. WiGR-AF uses passive sensing technology and utilizes existing Wi-Fi devices without the need for test personnel to carry specialized equipment. It extracts spatial and temporal information from the periodicity of gait using independent dual branches and designs an AFF-GAM attention mechanism to fuse the two types of gait features as a person's identity, resulting in more accurate gait representation, thus achieving direction and path independence. To verify the superiority of the method, we conducted extensive experiments on the public dataset Widar3.0, and the experimental results show that the method achieves an average recognition accuracy of 97.2% for 10 users, outperforming the baseline methods.