SVDNet for Pedestrian Retrieval

Citation:
2017
Issue Date:
2017
Full metadata record
Files in This Item:
Filename Description Size
1703.05693.pdfPublished version1.93 MB
Adobe PDF
SVDNet for Pedestrian Retrieval Yifan Sun y , Liang Zheng z , Weijian Deng x , Shengjin Wang y y Tsinghua University z University of Technology Sydney x University of Chinese Academy of Sciences sunyf15@mails.tsinghua.edu.cn, f liangzheng06, dengwj16 g @gmail.com, wgsgj@tsinghua.edu.cn Abstract This paper proposes the SVDNet for retrieval problems, with focus on the application of person re-identification (re- ID). We view each weight vector within a fully connected (FC) layer in a convolutional neuron network (CNN) as a projection basis. It is observed that the weight vectors are usually highly correlated. This problem leads to correla- tions among entries of the FC descriptor, and compromises the retrieval performance based on the Euclidean distance. To address the problem, this paper proposes to optimize the deep representation learning process with Singular Vector Decomposition (SVD). Specifically, with the restraint and relaxation iteration (RRI) training scheme, we are able to iteratively integrate the orthogonality constraint in CNN training, yielding the so-called SVDNet. We conduct ex- periments on the Market-1501, CUHK03, and DukeMTMC- reID datasets, and show that RRI effectively reduces the correlation among the projection vectors, produces more discriminative FC descriptors, and significantly improves the re-ID accuracy. On the Market-1501 dataset, for in- stance, rank-1 accuracy is improved from 55.3% to 80.5% for CaffeNet, and from 73.8% to 82.3% for ResNet-50.
Please use this identifier to cite or link to this item: