Specialized re-ranking: A novel retrieval-verification framework for cloth changing person re-identification
作者：
Highlights：
• We propose to use specialized features for discerning similar images.
• We propose a retrieval-verification framework for learning specialized features.
• A well designed verification network is introduced for comparing similar images.
• Our proposed ranking strategy can best exert the potential of verification results.
• We improve the state-of-the-art methods on both realistic and synthetic datasets.
摘要
•We propose to use specialized features for discerning similar images.•We propose a retrieval-verification framework for learning specialized features.•A well designed verification network is introduced for comparing similar images.•Our proposed ranking strategy can best exert the potential of verification results.•We improve the state-of-the-art methods on both realistic and synthetic datasets.
论文关键词：Cloth changing person re-identification,Verification network,Re-Rank,Specialized features,Part-based comparison
论文评审过程：Received 27 January 2022, Revised 5 August 2022, Accepted 21 September 2022, Available online 28 September 2022, Version of Record 20 October 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.109070
