Therefore, we used a spatial transformer network-based positioning model for the removal of background information in each of the two branches.<br><br>In this paper, the positioning model consists of three parts: the positioning network, the grid generator, and the bilinear sampler.
正在翻译中..