Person Re Identification On Dukemtmc Reid

评估指标

Rank-1

mAP

评测结果

各个模型在此基准测试上的表现结果

模型名称	Rank-1	mAP	Paper Title	Repository
Auto-ReID(RK)	91.4	89.2	Auto-ReID: Searching for a Part-aware ConvNet for Person Re-Identification
CTL Model (ResNet50, 256x128)	95.6	96.1	On the Unreasonable Effectiveness of Centroids in Image Retrieval
MAR	79.8	48	Unsupervised Person Re-identification by Soft Multilabel Learning
TriNet + Random Erasing	73.0	56.6	Random Erasing Data Augmentation
APR	70.69	51.88	Improving Person Re-identification by Attribute and Identity Learning
SSP-ReID	81.8	68.6	Improved Person Re-Identification Based on Saliency and Semantic Parsing with Deep Neural Network Models
OIM	68.1	47.4	Joint Detection and Identification Feature Learning for Person Search
DAAF-BoT	87.9	77.9	Deep Attention Aware Feature Learning for Person Re-Identification
Adaptive L2 Regularization (with re-ranking)	92.2	90.7	Adaptive L2 Regularization in Person Re-Identification
FlipReID (with re-ranking)	93.0	90.7	FlipReID: Closing the Gap between Training and Inference in Person Re-Identification
Top-DB-Net + RK	90.9	88.6	Top-DB-Net: Top DropBlock for Activation Enhancement in Person Re-Identification
ISP	89.6	80	Identity-Guided Human Semantic Parsing for Person Re-Identification
Pyramid (CVPR'19)	89.0	79.0	Pyramidal Person Re-IDentification via Multi-Loss Dynamic Training
DAAF-BoT(RK)	91.7	89.6	Deep Attention Aware Feature Learning for Person Re-Identification
GAN	67.68	47.13	Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro
DG-Net(RK)	90.26	88.31	Joint Discriminative and Generative Learning for Person Re-identification
Incremental Learning	80.0	60.2	Incremental Learning in Person Re-Identification
RPTM	93.5	89.2	Relation Preserving Triplet Mining for Stabilising the Triplet Loss in Re-identification Systems
LDS (ResNet50 + RK)	92.91	91.0	Learning to Disentangle Scenes for Person Re-identification	-
Deep Miner (w/o ReRank)	91.20	81.80	Deep Miner: A Deep and Multi-branch Network which Mines Rich and Diverse Features for Person Re-identification

0 of 94 row(s) selected.