[PersonReID] Part-Aligned Representations for Person Re-Identification

ICCV2017 paper: Deeply-Learned Part-Aligned Representations for Person Re-Identification code caffe 網絡結構代碼 caffe網絡結構可視化 論文中的網絡結構 拿出inception的最後一個卷積層的feature map的輸出：（ w ,h ,512) 1.輸入feature map ，用1x1的卷

>>阅读原文<<