Discrete online cross-modal hashing

作者：

Highlights：

• Different from the majority of related methods, DOCH is a discrete one.

• By keeping the binary constraints, quantization error can be avoided.

• By preserving the similarity in Hamming space, DOCH learns accurate hash codes.

• Both time and space complexity are low and acceptable.

• Extensive experiments demonstrate the superiority of DOCH.

摘要

•Different from the majority of related methods, DOCH is a discrete one.•By keeping the binary constraints, quantization error can be avoided.•By preserving the similarity in Hamming space, DOCH learns accurate hash codes.•Both time and space complexity are low and acceptable.•Extensive experiments demonstrate the superiority of DOCH.

论文关键词：Cross-modal retrieval,Discrete optimization,Online hashing,Learning to hash

论文评审过程：Received 19 June 2021, Revised 12 August 2021, Accepted 18 August 2021, Available online 19 August 2021, Version of Record 3 September 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108262