Self-paced Contrastive Learning with Hybrid Memory for Domain Adaptive Object Re-ID
Yixiao Ge Feng Zhu Dapeng Chen Rui Zhao Hongsheng Li Multimedia Laboratory The Chinese University of Hong Kong
Abstract
Domain adaptive object re-ID aims to transfer the learned knowledge from the labeled source domain to the unlabeled target domain to tackle the open-class re-identification problems. Although state-of-the-art pseudo-label-based methods [11, 54, 10, 55, 14] have achieved great success, they did not make full use of all valuable information because of the domain gap and unsatisfying clustering performance. To solve these problems, we propose a novel self-paced contrastive learning framework with hybrid memory. The hybrid memory dynamically generates source-domain class-level, target-domain cluster-level and un-clustered instance-level supervisory signals for learning feature representations. Different from the conventional contrastive learning strategy, the proposed framework jointly distinguishes source-domain classes, and target-domain clusters and un-clustered instances. Most importantly, the proposed self-paced method gradually creates more reliable clusters to refine the hybrid memory and learning targets, and is shown to be the key to our outstanding performance. Our method outperforms state-of-the-arts on multiple domain adaptation tasks of object re-ID and even boosts the per
中文速览
跨摄像头、跨场景的行人/车辆重识别（object re-ID）在实际落地时常常面临"有标注的源域数据"和"无标注的目标域数据"之间的巨大鸿沟，现有方法要么直接丢弃源域标签、要么扔掉聚类时产生的离群点，白白浪费了大量有价值的监督信息。针对这一痛点，作者提出了一套"混合记忆+自步对比学习（Self-paced Contrastive Learning with Hybrid Memory，SpCL）"框架：用一个动态记忆库同时存储源域类别中心、目标域聚类中心和未被聚类的孤立样本特征，三类信息共同驱动一个统一的对比损失来学习判别性特征；与此同时，"自步策略"从最可靠的聚类出发，随着训练推进逐步把更多离群点纳入新的可靠聚类，有效抑制了噪声伪标签带来的误差累积。实验结果表明，该方法在多个域自适应行人/车辆重识别基准上比此前最优方法提升最高达5.0% mAP，在完全无监督设置下于Market-1501和MSMT17两个数据集上分别领先16.7%和7.9%，意义在于它首次把源域标签、目标域聚类信息和离群样本三者统一纳入对比学习，为跨域重识别问题提供了一个更充分利用所有可用数据的通用范式。
原文 arXiv:2006.02713;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2006.02713v2