View Confusion Feature Learning for Person Re-identification

Author Fangyi Liu Conference ICCV 2019 Motivation ReID的性能受图片拍摄视角影响,所以希望提取view-invariant特征。 Contribution 使用view confusion learning mechanism,提取视角不变性特征。 用SIFT特征来指导模型训练。 Framework Pre-trained CNN Network
相关文章
相关标签/搜索