Hard-sample Guided Hybrid Contrast Learning for Unsupervised Person Re-Identification

25 Sep 2021  ·  Zheng Hu, Chuang Zhu, Gang He ·

Unsupervised person re-identification (Re-ID) is a promising and very challenging research problem in computer vision. Learning robust and discriminative features with unlabeled data is of central importance to Re-ID. Recently, more attention has been paid to unsupervised Re-ID algorithms based on clustered pseudo-label. However, the previous approaches did not fully exploit information of hard samples, simply using cluster centroid or all instances for contrastive learning. In this paper, we propose a Hard-sample Guided Hybrid Contrast Learning (HHCL) approach combining cluster-level loss with instance-level loss for unsupervised person Re-ID. Our approach applies cluster centroid contrastive loss to ensure that the network is updated in a more stable way. Meanwhile, introduction of a hard instance contrastive loss further mines the discriminative information. Extensive experiments on two popular large-scale Re-ID benchmarks demonstrate that our HHCL outperforms previous state-of-the-art methods and significantly improves the performance of unsupervised person Re-ID. The code of our work is available soon at https://github.com/bupt-ai-cz/HHCL-ReID.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Unsupervised Person Re-Identification DukeMTMC-reID HHCL(ResNet50 w/o RK) Rank-1 85.1 # 1
Rank-10 94.6 # 1
Rank-5 92.4 # 1
MAP 73.3 # 1
Unsupervised Person Re-Identification Market-1501 HHCL(ResNet50 w/o RK) Rank-1 93.4 # 3
MAP 84.2 # 2
Rank-10 98.5 # 1
Rank-5 97.7 # 1

Methods


No methods listed for this paper. Add relevant methods here