Three Stream Spatiotemporal Gait Neural Network for Kinect Identity Authentication
Gait recognition is a biometric recognition technology.To improve the accuracy of gait recognition in corridor scenes,this paper proposes a hybrid deep learning model using three stream graph structured data.The model is based on both temporal and spatial convolutional networks to fully extract the most representative spatiotemporal gait dynamic features from gait maps.Through experiments on normal gait data of 30 subjects,the results showed that the maximum accuracy of this method reached 99.58%,indicating that the method can learn spatiotemporal differences in gait features.
万方数据
维普
