ZTE Communications ›› 2022, Vol. 20 ›› Issue (1): 76-82.doi: 10.12142/ZTECOM.202201010

• Research Paper • Previous Articles    

Metric Learning for Semantic‑Based Clothes Retrieval

YANG Bo1, GUO Caili1,2(), LI Zheng1   

  1. 1.Beijing Laboratory of Advanced Information Networks, Beijing University of Posts and Telecommunications, Beijing 100876, China
    2.Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2021-11-15 Online:2022-03-25 Published:2022-04-06
  • About author:YANG Bo received the B.S. degree in communication engineering from Beijing University of Posts and Telecommunications (BUPT), China in 2019. He is currently pursuing the M.S. degree in information and communication engineering at BUPT. His current research interests include computer vision and image retrieval.|GUO Caili (guocaili@bupt.edu.cn) received the Ph.D. degree in communication and information systems from Beijing University of Posts and Telecommunication (BUPT), China in 2008. She is currently a professor in the School of Information and Communication Engineering, BUPT. Her general research interests include machine learning and statistical signal processing, with current emphasis on semantic communications, deep learning, and intelligence visual computing. In the related areas, she has published over 200 papers and holds over 30 granted patents. She won Diamond Best Paper Award of IEEE ICME 2018 and Best Paper Award of IEEE WCNC 2021.|LI Zheng received the B.S. degree in telecommunication engineering from Shandong University, China in 2016, and the M.S. degree in information and communication engineering from Beijing University of Posts and Telecommunications (BUPT), China in 2019. He is currently pursuing the Ph.D. degree in information and communication engineering at BUPT. His current research interests include computer vision and multimedia retrieval.

Abstract:

Existing clothes retrieval methods mostly adopt binary supervision in metric learning. For each iteration, only the clothes belonging to the same instance are positive samples, and all other clothes are “indistinguishable” negative samples, which causes the following problem. The relevance between the query and candidates is only treated as relevant or irrelevant, which makes the model difficult to learn the continuous semantic similarities between clothes. Clothes that do not belong to the same instance are completely considered irrelevant and are uniformly pushed away from the query by an equal margin in the embedding space, which is not consistent with the ideal retrieval results. Motivated by this, we propose a novel method called semantic-based clothes retrieval (SCR). In SCR, we measure the semantic similarities between clothes and design a new adaptive loss based on these similarities. The margin in the proposed adaptive loss can vary with different semantic similarities between the anchor and negative samples. In this way, more coherent embedding space can be learned, where candidates with higher semantic similarities are mapped closer to the query than those with lower ones. We use Recall@K and normalized Discounted Cumulative Gain (nDCG) as evaluation metrics to conduct experiments on the DeepFashion dataset and have achieved better performance.

Key words: clothes retrieval, metric learning, semantic-based retrieval