High-Accuracy Gesture Recognition using Mm-Wave Radar Based on Convolutional Block Attention Module

Non-contact gesture recognition is a new type of human-computer interaction with broad prospects in many applications. Motivated by the need for more precise micro-motion gesture recognition using mm-wave radar in recent years, a novel micro-motion gesture recognition network based on the Convolutional Block Attention Module (CBAM) is proposed here. The MMWCAS radar from TI is used to collect gesture echoes. During data pre-processing, the Range-time Map, Doppler-time Map, Azimuth-time Map and Elevation-time Map of the gestures are extracted and employed to characterize the motion features. A DenseNet and CBAM-based gesture recognition network is designed to identify the 12 types of micro-motion gestures using the mixed Feature-time Map as input. According to the experimental results, the accuracy rate reaches 99.03%, achieving high-accuracy gesture recognition. It has been discovered that the network focuses on the first half of the gesture movement, which improves recognition accuracy.