Affordable Access

Access to the full text

A Comprehensive Study on Center Loss for Deep Face Recognition

Authors
  • Wen, Yandong1
  • Zhang, Kaipeng1
  • Li, Zhifeng2
  • Qiao, Yu3
  • 1 Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen Key Lab on Computer Vision and Pattern Recognition, Shenzhen, China , Shenzhen (China)
  • 2 Tencent AI Lab, Shenzhen, China , Shenzhen (China)
  • 3 SIAT-SenseTime Joint Lab, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China , Shenzhen (China)
Type
Published Article
Journal
International Journal of Computer Vision
Publisher
Springer-Verlag
Publication Date
Jan 17, 2019
Volume
127
Issue
6-7
Pages
668–683
Identifiers
DOI: 10.1007/s11263-018-01142-4
Source
Springer Nature
Keywords
License
Yellow

Abstract

Deep convolutional neural networks (CNNs) trained with the softmax loss have achieved remarkable successes in a number of close-set recognition problems, e.g. object recognition, action recognition, etc. Unlike these close-set tasks, face recognition is an open-set problem where the testing classes (persons) are usually different from those in training. This paper addresses the open-set property of face recognition by developing the center loss. Specifically, the center loss simultaneously learns a center for each class, and penalizes the distances between the deep features of the face images and their corresponding class centers. Training with the center loss enables CNNs to extract the deep features with two desirable properties: inter-class separability and intra-class compactness. In addition, we extend the center loss in two aspects. First, we adopt parameter sharing between the softmax loss and the center loss, to reduce the extra parameters introduced by centers. Second, we generalize the concept of center from a single point to a region in embedding space, which further allows us to account for intra-class variations. The advanced center loss significantly enhances the discriminative power of deep features. Experimental results show that our method achieves high accuracies on several important face recognition benchmarks, including Labeled Faces in the Wild, YouTube Faces, IJB-A Janus, and MegaFace Challenging 1.

Report this publication

Statistics

Seen <100 times