Hand gesture recognition method based on dual-channel convolutional neural network

Abstract

With the rapid development of security monitoring, assisted driving, remote health diagnosis and other fields in recent years, the recognition of human characteristics has attracted more and more attention. The wideband radar has a high range resolution compared to the narrow-band radar which make it able to extract fine features of human micro-movements. However, the micro-movement features of the human body are often in a complex background. Furthermore, the micro-movement features of the human body are weak compared to the main body. Therefore, the classification and recognition of human micro-motion based on wideband radar is still a difficult problem. Inspired by the successful application of convolutional neural network in image processing, this paper proposes a wideband hand gesture recognition method based on dual-channel convolutional neural network for wideband radar, which takes the range-Doppler map and high resolution range profile of human micro-motion as inputs. The effectiveness of this method is verified by experimental data, after the information is convolved, the features are fused, and finally the purpose of classification is achieved. The target recognition rate of this method is 95.67%, which is much higher than 89.87% of the High Resolution Range Profile(HRRP) and 88.61% of the Range Doppler(RD), which verifies the effectiveness of the method.