Camera Style Adaptation for Person Re-identification

CVPR 2018 Zhun ZhongLiang ZhengZhedong ZhengShaozi LiYi Yang

Being a cross-camera retrieval task, person re-identification suffers from image style variations caused by different cameras. The art implicitly addresses this problem by learning a camera-invariant descriptor subspace... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Person Re-Identification DukeMTMC-reID IDE* + CamStyle + Random Erasing Rank-1 78.32 # 17
Person Re-Identification DukeMTMC-reID IDE* + CamStyle + Random Erasing MAP 57.61 # 21
Person Re-Identification DukeMTMC-reID IDE* + CamStyle Rank-1 75.27 # 19
Person Re-Identification DukeMTMC-reID IDE* + CamStyle MAP 53.48 # 25
Person Re-Identification DukeMTMC-reID IDE* Rank-1 72.31 # 22
Person Re-Identification DukeMTMC-reID IDE* MAP 51.83 # 27
Person Re-Identification Market-1501 IDE* + CamStyle + Random Erasing Rank-1 89.49 # 20
Person Re-Identification Market-1501 IDE* + CamStyle + Random Erasing MAP 71.55 # 23
Person Re-Identification Market-1501 IDE* + CamStyle Rank-1 88.12 # 23
Person Re-Identification Market-1501 IDE* + CamStyle MAP 68.72 # 28
Person Re-Identification Market-1501 IDE* Rank-1 85.66 # 25
Person Re-Identification Market-1501 IDE* MAP 65.87 # 30