【ReID】Learning Discriminative Features with Multiple Granularities for Person Re-identification

閱讀2018 ACM Multimedia一篇著名SOTA,Learning Discriminative Features with Multiple Granularities for Person Re-Identification[1],文章設計了一個結合多分支和表徵度量雙學習策略的切片網絡,提取全局粗粒度和局部細粒度特徵。通過多種細粒度特徵的提取和學習來得到更好的具有判別力的表徵,實驗顯
相關文章
相關標籤/搜索