Circle Loss: A Unified Perspective of Pair Similarity Optimization

• Published in: Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) pp. 6397 - 6406
• Main Authors: Sun, Yifan, Cheng, Changmao, Zhang, Yuhan, Zhang, Chi, Zheng, Liang, Wang, Zhongdao, Wei, Yichen
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.01.2020
• Subjects: Convergence; Face recognition; Image retrieval; Measurement; Optimization; Task analysis; Training
• ISSN: 1063-6919
• DOI: 10.1109/CVPR42600.2020.00643

This paper provides a pair similarity optimization viewpoint on deep feature learning, aiming to maximize the within-class similarity s_p and minimize the between-class similarity s_n. We find a majority of loss functions, including the triplet loss and the softmax cross-entropy loss, embed s_n and s_p into similarity pairs and seek to reduce (s_n-s_p). Such an optimization manner is inflexible, because the penalty strength on every single similarity score is restricted to be equal. Our intuition is that if a similarity score deviates far from the optimum, it should be emphasized. To this end, we simply re-weight each similarity to highlight the less-optimized similarity scores. It results in a Circle loss, which is named due to its circular decision boundary. The Circle loss has a unified formula for two elemental deep feature learning paradigms, \emph {i.e.}, learning with class-level labels and pair-wise labels. Analytically, we show that the Circle loss offers a more flexible optimization approach towards a more definite convergence target, compared with the loss functions optimizing (s_n-s_p). Experimentally, we demonstrate the superiority of the Circle loss on a variety of deep feature learning tasks. On face recognition, person re-identification, as well as several fine-grained image retrieval datasets, the achieved performance is on par with the state of the art.