Gesture Recognition System based on Millimeter Wave and Thermal Imager
Wen-Hsiang Yeh, Yi‐Lin Cheng, Yu‐Ping Liao · 2023
With the COVID-19 outbreak, the demand for replacing traditional button control or touch screens is gradually increasing. However, most gesture recognition technologies rely on machine vision methods only. Poor recognition results are yielded with the camera in weak light or with complex backgrounds. Thus, we proposed a large-motion gesture recognition system that combined millimeter-wave radar and thermal imager, and deep learning for improved accuracy. As the user gestured, point cloud information was captured and analyzed by a neural network model. The system also used thermal imaging and palm recognition to track hand movements across the screen. The results showed that this integration improves accuracy to over 80%.