Cross-domain gait re-identification method based on optimized clustering algorithm

The invention provides a cross-domain gait re-identification method based on an optimized clustering algorithm. The method comprises the following steps: firstly, initializing GaitSet network parameters by using a labeled source domain training set; secondly, clustering and assigning a pseudo label to a label-free target domain training set through an optimized clustering algorithm by utilizing the characteristics output by the trained GaitSet network, so as to obtain a target domain training set with the pseudo label; finally, updating GaitSet network parameters through the target domain training set with the pseudo label yt; finally, carrying out cross-domain gait re-identification through the updated GaitSet network. According to the method, the problems of unreliability and large noiseof a clustering result of a traditional clustering algorithm are solved, the reliability of the clustering result is improved, and the recognition precision in a cross-domain scene is improved. 本发明提供一种基于优化聚类算法的跨域步态重识别方法，首先使用