A Novel Robust Method for Hand Gesture Recognition Based on Millimeter-Wave Radar
Hongzhi Li, Lanyang Kang, Qinglong Hua, Bin Zhao · 2025
As one of the most commonly used communication methods, gestures have gradually become a research hotspot in human-computer interaction. To address the issue of target-like interference in millimeter-wave radar gesture recognition, this paper will employ the Cubature Kalman Filter (CKF) algorithm to process gesture data and compare the results with those obtained by the Unscented Kalman Filter (UKF) algorithm. To tackle the problems of low accuracy and poor robustness in gesture recognition using deep learning, this paper constructs a Dual-Stream Fusion Residual Network (DSFRN) by leveraging multi-feature domains after feature fusion for gesture data learning and recognition. Then we integrate the Convolutional Block Attention Module (CBAM) into the network. Experiments show that the recognition accuracy of fused features can reach 97.45%.