WiHF: Enable User Identified Gesture Recognition with WiFi

Abstract

User identified gesture recognition is a fundamental step towards ubiquitous device-free sensing. We propose WiHF, which first simultaneously enables cross-domain gesture recognition and user identification using WiFi in a real-time manner. The basic idea of WiHF is to derive a cross-domain motion change pattern of arm gestures from WiFi signals, rendering both unique gesture characteristics and the personalized user performing styles. To extract the motion change pattern in realtime, we develop an efficient method based on the seam carving algorithm. Moreover, taking as input the motion change pattern, a Deep Neural Network (DNN) is adopted for both gesture recognition and user identification tasks. In DNN, we apply splitting and splicing schemes to optimize collaborative learning for dual tasks. We implement WiHF and extensively evaluate its performance on a public dataset including 6 users and 6 gestures performed across 5 locations and 5 orientations in 3 environments. Experimental results show that WiHF achieves 97.65% and 96.74% for in-domain gesture recognition and user identification accuracy, respectively. The cross-domain gesture recognition accuracy is comparable with the state-of-the-art methods, but the processing time is reduced by 30×.