Self-paced Contrastive Learning with Hybrid Memory for Domain Adaptive Object Re-ID

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 have achieved great success, they did not make full use of all valuable information because of the domain gap and unsatisfying clustering performance... (read more)

PDF Abstract NeurIPS 2020 PDF NeurIPS 2020 Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Unsupervised Domain Adaptation Duke to Market SpCL mAP 76.7 # 2
rank-1 90.3 # 2
rank-5 96.2 # 2
rank-10 97.7 # 2
Unsupervised Domain Adaptation Duke to MSMT SpCL mAP 26.5 # 2
rank-1 53.1 # 3
rank-5 65.8 # 3
rank-10 70.5 # 3
Unsupervised Domain Adaptation Market to Duke SpCL mAP 68.8 # 2
rank-1 82.9 # 2
rank-5 90.1 # 2
rank-10 92.5 # 2
Unsupervised Domain Adaptation Market to MSMT SpCl mAP 25.4 # 2
rank-1 51.6 # 2
rank-5 64.3 # 2
rank-10 69.7 # 2

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet