In this paper, stargan is used to learn the transformation model of camera style from the source domain to each target camera.<br>Then, the source image style of each marker is transferred to the target camera, and the generated image is used to train the re ID model of the target domain through the proposed feature learning network.<br>In this way, our method not only reduces the gap between the source domain and the target domain, but also considers the changes in the target domain.<br>
正在翻译中..