Deep center-based dual-constrained hashing for discriminative face image retrieval

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

24 Scopus Citations
View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageEnglish
Article number107976
Journal / PublicationPattern Recognition
Volume117
Online published6 Apr 2021
Publication statusPublished - Sept 2021

Abstract

With the advantages of low storage cost and extremely fast retrieval speed, deep hashing methods have attracted much attention for image retrieval recently. However, large-scale face image retrieval with significant intra-class variations is still challenging. Neither existing pairwise/triplet labels-based nor softmax classification loss-based deep hashing works can generate compact and discriminative binary codes. Considering these issues, we propose a center-based framework integrating end-to-end hashing learning and class centers learning simultaneously. The framework minimizes the intra-class variance by clustering intra-class samples into a learnable class center. To strengthen inter-class separability, it additionally imposes a novel regularization term to enlarge the Hamming distance between pairwise class centers. Moreover, a simple yet effective regression matrix is introduced to encourage intra-class samples to generate the same binary codes, which further enhances the hashing codes compactness. Experiments on four large-scale datasets show the proposed method outperforms state-of-the-art baselines under various code lengths and commonly-used evaluation metrics.

Research Area(s)

  • Class centers, Convolutional neural networks, Deep supervised hashing, Face image retrieval