Based on the above issues, we propose a graph-based soft fusion method that integrates spatial-temporal information. By limiting the scope of the spatial-temporal information and only using it as a means of selecting neighbors, we reduce the impact of noise introduced by fusing spatial-temporal information. This approach is tailored to the field of pedestrian re-identification.

帮我把下面这段话按照行人重识别领域的论文的风格翻译成英文:基于上述问题我们提出了一种基于图的融合时空信息的软结合方法通过限制时空信息的作用范围并且仅将时空信息作为筛选邻居的方式来降低融合时空所引入的噪声的影响。

原文地址: https://www.cveoy.top/t/topic/cq0B 著作权归作者所有。请勿转载和采集!

免费AI点我,无需注册和登录