Work in Progress: Enabling User Identification for mmWave-based Gesture Recognition Systems
Lilin Xu, K.R. Wang, Chaojie Gu, Shibo He, Jiming Chen · 2023
The mmWave radar has been exploited for gesture recognition. However, existing mmWave-based gesture recognition methods cannot identify different users, which is important for ubiquitous gesture interaction in many applications. This paper proposes GesturePrint, which is the first to achieve person-independent gesture recognition and gesture-based user identification using a commodity mmWave radar sensor. GesturePrint features an effective pipeline that enables the gesture recognition system to identify users with a minor additional cost. By introducing an efficient signal preprocessing stage and a novel network architecture GesIDNet, which employs an attention-based adaptive multilevel feature fusion mechanism, GesturePrint extracts both unique characteristics of predefined gestures and effective features of personalized motion patterns. Experiments on our self-collected dataset and three public datasets demonstrate GesturePrint's superior performance in enabling user identification for gesture recognition systems.