Skip to main content

Advertisement

Table 2 Comparison with the state-of-the-art methods on the CUHK03 dataset under the single-shot setting

From: Learning deep features from body and parts for person re-identification in camera networks

Method Rank-1 Rank-5 Rank-10 mAP
KISSME [38] 11.7 33.3 48.0
DeepReID [39] 19.9 49.3 64.7
BoW+HS [19] 24.3
LOMO+XQDA [7] 46.3 78.9 88.6
SI-CI [40] 52.2 84.3 94.8
DNS [32] 54.7 80.1 88.3
SOMAnet [35] 72.4 92.1 95.8
Part-based features 17.8 46.7 64.3 24.8
Body-based features 74.2 92.5 96.8 78.3
DFBP 75.9 93.8 97.9 79.9
  1. The data in italics are the best result in each evaluation protocol