HyperAI超神经
首页资讯最新论文教程数据集百科SOTALLM 模型天梯GPU 天梯顶会
全站搜索
关于
HyperAI超神经

Person Re Identification On Cuhk03 Labeled

评估指标

MAP
Rank-1

评测结果

各个模型在此基准测试上的表现结果

模型名称
MAP
Rank-1
Paper TitleRepository
IDE-C+XQDA20.021.9Re-ranking Person Re-identification with k-reciprocal Encoding-
MGN (ACM MM'18)67.468.0Learning Discriminative Features with Multiple Granularities for Person Re-Identification
Weakly Pre-training (ResNet101+RK)89.2987.86--
LightMBN (w/o ReRank)85.187.2Lightweight Multi-Branch Network for Person Re-Identification
PAN+re-rank45.843.9Pedestrian Alignment Network for Large-scale Person Re-identification
Auto-ReID (ICCV'19)73.077.9Auto-ReID: Searching for a Part-aware ConvNet for Person Re-Identification
IDE-C14.915.6Re-ranking Person Re-identification with k-reciprocal Encoding-
IDE-R+XQDA29.632.0Re-ranking Person Re-identification with k-reciprocal Encoding-
MPN (without re-ranking)81.185Multi-task Learning with Coarse Priors for Robust Part-aware Person Re-identification
DaRe+RE (CVPR'18)61.666.1Resource Aware Person Re-identification across Multiple Resolutions
HA-CNN (CVPR'18)41.044.4Harmonious Attention Network for Person Re-Identification
Deep Miner (w/o ReRank)84.786.6Deep Miner: A Deep and Multi-branch Network which Mines Rich and Diverse Features for Person Re-identification
Top-DB-Net75.479.4Top-DB-Net: Top DropBlock for Activation Enhancement in Person Re-Identification
IDE-R21.022.2Re-ranking Person Re-identification with k-reciprocal Encoding-
Pyramid (CVPR' 19)76.978.9Pyramidal Person Re-IDentification via Multi-Loss Dynamic Training
DiP (without RK)85.787DiP: Learning Discriminative Implicit Parts for Person Re-Identification
PLR-OSNet80.584.6Learning Diverse Features with Part-Level Resolution for Person Re-Identification
FPB83.885.9FPB: Feature Pyramid Branch for Person Re-Identification
Top-DB-Net + RK88.586.7Top-DB-Net: Top DropBlock for Activation Enhancement in Person Re-Identification
PAN(Zheng et al., [2017a])35.036.9Pedestrian Alignment Network for Large-scale Person Re-identification
0 of 21 row(s) selected.