Semantic-rebased cross-modal hashing for scalable unsupervised text-visual retrieval

作者：

Highlights：

• A semantic rebasing mechanism is proposed to learn a sparse semantic graph structure to preserve similarity information of different modalities for binary code learning.

• The proposed method focuses on both similarity preserving and quantization to gain satisfied retrieval performance.

• Auto-encoding structure of both modalities is included to improve the performance.

• Extensive experiments on four popular datasets shows the effectiveness of the proposed method.

摘要

•A semantic rebasing mechanism is proposed to learn a sparse semantic graph structure to preserve similarity information of different modalities for binary code learning.•The proposed method focuses on both similarity preserving and quantization to gain satisfied retrieval performance.•Auto-encoding structure of both modalities is included to improve the performance.•Extensive experiments on four popular datasets shows the effectiveness of the proposed method.

论文关键词：Sparse graph,Semantic rebasing,Cross-modal hashing,Unsupervised text-visual retrieval

论文评审过程：Received 31 May 2020, Revised 25 July 2020, Accepted 15 August 2020, Available online 20 August 2020, Version of Record 20 October 2020.

论文官网地址：https://doi.org/10.1016/j.ipm.2020.102374