Hand Gesture Recognition Using IR-UWB Radar with ShuffleNet V2
Yao Li, Xin Wang, Baodai Shi, Mingming Zhu · 2021
In recent years, gesture recognition has developed rapidly in non-contact Human-Computer Interaction(HCI).This paper presents an efficient hand gesture recognition system for HCI based on Impulse-Radio Ultra-Wideband (IR-UWB) radar using ShuffleNet V2, which performed well on accuracy, speed and robustness. We convert time-domain radar signals to continuous Range-Doppler Map(RDM) by algorithm1. RDM images are easier to understand than waveform diagrams for Convolutional Neural Networks(CN-N). ShuffleNet V2 is a masterpiece of lightweight CNN and is used to analyze the patterns of different gesture RDM images to classify gestures in this paper. In order to ensure the robustness of the algorithm, we invited 7 participants to construct the gesture data set. The proposed hand gesture recognition system can classify 7 gestures with a promising accuracy of 98.52%.