Deep momentum uncertainty hashing
作者：
Highlights：
• We are the first to explore the uncertainty of hashing bits during approximate optimization. Depending on the magnitude of uncertainty, the corresponding hashing bits and input images receive different attention.
• We propose to explicitly model bit-level and image-level uncertainty, resorting to the output discrepancy between the hashing network and the momentum-updated network.
• Extensive experiments on the CIFAR-10, the NUS-WIDE, the MS-COCO, and the largescale Clothing1M datasets demonstrate that our method significantly improves retrieval performance when compared with state-of-the-art methods.
摘要
•We are the first to explore the uncertainty of hashing bits during approximate optimization. Depending on the magnitude of uncertainty, the corresponding hashing bits and input images receive different attention.•We propose to explicitly model bit-level and image-level uncertainty, resorting to the output discrepancy between the hashing network and the momentum-updated network.•Extensive experiments on the CIFAR-10, the NUS-WIDE, the MS-COCO, and the largescale Clothing1M datasets demonstrate that our method significantly improves retrieval performance when compared with state-of-the-art methods.
论文关键词：Combinatorial optimization,Deep hashing,Uncertainty
论文评审过程：Received 7 January 2021, Revised 16 August 2021, Accepted 18 August 2021, Available online 19 August 2021, Version of Record 26 August 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108264
