Digital Gesture Recognition Based on Millimeter Wave Radar

Abstract

Gesture recognition technology is an effective way of natural and intuitive communication between humans and machines. Among the many existing recognition devices, millimeter wave radar has become a new approach to gesture recognition technology because of its strong target detection capability, insensitivity to changes in factors such as light and environment, low operating energy consumption and protection of user privacy. In this paper, we propose an over-the-air handwritten digit recognition method based on a millimeter wave radar. Unlike the direct use of spectrograms for gesture recognition, we use the trajectory points of gesture actions to generate images for gesture recognition. In this paper, a signal processing process and an image conversion method are proposed for generating gesture recognition images. Finally, to demonstrate the accuracy and effectiveness of the processing method, we conducted experiments on handwritten digits and used Xception as the base network for image recognition with an accuracy of 99.6%, proving the performance of the method.