Attentive WaveBlock: Complementarity-enhanced Mutual Networks for Unsupervised Domain Adaptation in Person Re-identification and Beyond

11 Jun 2020 Wenhao Wang Fang Zhao Shengcai Liao Ling Shao

Unsupervised domain adaptation (UDA) for person re-identification is challenging because of the huge gap between the source and target domain. A typical self-training method is to use pseudo-labels generated by clustering algorithms to iteratively optimize the model on the target domain... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Unsupervised Domain Adaptation Duke to Market AWB mAP 80.6 # 1
rank-1 92.9 # 1
rank-5 97.2 # 1
rank-10 98.2 # 1
Unsupervised Domain Adaptation Duke to MSMT AWB mAP 30.7 # 1
rank-1 62.7 # 1
rank-5 74.5 # 1
rank-10 79.0 # 1
Unsupervised Domain Adaptation Market to Duke AWB mAP 71.0 # 1
rank-1 83.4 # 1
rank-5 91.7 # 1
rank-10 93.8 # 1
Unsupervised Domain Adaptation Market to MSMT AWB mAP 30.6 # 1
rank-1 61.4 # 1
rank-5 73.3 # 1
rank-10 78.2 # 1

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