paper-with-me

Person Re-Identification 벤치마크

Person Re-Identification on DukeMTMC-reID

94개 결과 · ⬇ CSV · JSON

mAP

12.17 33.4 54.63 75.87 97.1 2014-06 2026-09 LOMO + XQDA — 17.04 (2014-06-17) BOW — 12.17 (2015-12-01) OIM — 47.4 (2016-04-07) IDE — 44.99 (2016-10-10) DLCE — 49.3 (2016-11-17) GAN — 47.13 (2017-01-26) SVDNet — 56.8 (2017-03-16) APR — 51.88 (2017-03-21) TriNet — 53.5 (2017-03-22) PUL* — 16.4 (2017-05-30) PointNet++ (MSG) [qi2017pointnet++] — 39.36 (2017-06-07) PAN + re-rank — 66.74 (2017-07-03) PAN — 51.51 (2017-07-03) ShuffleNetV2 [zhang2018shufflenet] — 48.09 (2017-07-04) SVDNet + Random Erasing — 62.4 (2017-08-16) TriNet + Random Erasing — 56.6 (2017-08-16) SPGAN+LMP* — 26.2 (2017-11-19) PCB (RPP) — 69.2 (2017-11-26) PCB (UP) — 66.1 (2017-11-26) PSE+ ECN (rank-dist) — 79.8 (2017-11-28) IDE* + CamStyle + Random Erasing — 57.61 (2017-11-28) IDE* — 51.83 (2017-11-28) MGN — 78.4 (2018-04-04) HPM — 74.3 (2018-04-14) DaRe(De)+RE+RR [wang2018resource] — 80.0 (2018-05-22) SSP-ReID (RR) — 83.7 (2018-07-15) SSP-ReID — 68.6 (2018-07-15) Proposed SGGNN — 68.2 (2018-07-26) Incremental Learning — 60.2 (2018-08-20) FD-GAN — 64.5 (2018-10-06) Pyramid (CVPR'19) — 79.0 (2018-10-29) Parameter-Free Spatial Attention — 85.9 (2018-11-29) st-ReID(RE, RK,Cam) — 92.7 (2018-12-08) UTAL — 44.6 (2019-03-01) MAR — 48.0 (2019-03-15) BoT Baseline(RK) — 89.1 (2019-03-17) Auto-ReID(RK) — 89.2 (2019-03-23) Bias-controlled Adversarial Training — 74.8 (2019-03-30) DG-Net(RK) — 88.31 (2019-04-15) DG-Net — 74.8 (2019-04-15) Deep Constrained Dominant Sets — 86.1 (2019-04-25) OSNet (ICCV'19) — 73.5 (2019-05-02) Dispersion based Clustering — 30.0 (2019-06-04) ReID strong baseline (IBN-Net50-a) — 79.1 (2019-06-19) PyrNet (+ReRank) — 87.7 (2019-06-20) PyrNet — 74.0 (2019-06-20) ABD-Net (ResNet-50) — 78.59 (2019-08-03) Compact-ReID (6.4M w/o RK) — 80.3 (2019-10-15) Compact-ReID (2.9M w/o RK) — 78.9 (2019-10-15) P2-Net (triplet loss) — 73.1 (2019-10-22) IS-GAN — 79.5 (2019-10-26) Viewpoint-Aware Loss(RK) — 91.8 (2019-12-03) PLR-OSNet — 81.2 (2020-01-21) CBN+BoT* — 70.1 (2020-01-23) CBN — 67.3 (2020-01-23) DAAF-BoT(RK) — 89.6 (2020-03-01) DAAF-BoT — 77.9 (2020-03-01) MPN (without re-ranking) — 82.0 (2020-03-18) OGNet — 57.89 (2020-06-08) Adaptive L2 Regularization (with re-ranking) — 90.7 (2020-07-15) Adaptive L2 Regularization (without re-ranking) — 81.0 (2020-07-15) ISP — 80.0 (2020-07-27) Cluster-level Alignment (Resnet50 w/o RK) — 81.84 (2020-08-15) CACENET (ResNet50 w/o RK) — 81.29 (2020-09-11) Top-DB-Net + RK — 88.6 (2020-10-12) Top-DB-Net — 73.5 (2020-10-12) Unsupervised Pre-training (ResNet101+RK) — 92.77 (2020-12-07) Unsupervised Pre-training (ResNet101+MGN) — 84.1 (2020-12-07) CAP — 76.0 (2020-12-19) RGT&RGPR (RK) — 92.7 (2021-01-21) TransReID (w/o RK) — 82.1 (2021-02-08) Deep Miner (w/o ReRank) — 81.8 (2021-02-18) + Re-weighting — 79.1 (2021-03-02) DenseIL — 97.1 (2021-03-16) GPS (without re-ranking) — 78.7 (2021-04-14) CTL Model (ResNet50, 256x128) — 96.1 (2021-04-28) FlipReID (with re-ranking) — 90.7 (2021-05-12) FlipReID (without re-ranking) — 81.5 (2021-05-12) st-ReID+InSTD — 89.1 (2021-07-31) FPB — 82.9 (2021-08-04) APNet-C — 81.5 (2021-08-11) CAL — 80.5 (2021-08-19) RPTM — 89.2 (2021-10-15) LDS (ResNet50 + RK) — 91.0 (2021-11-10) Weakly Supervised Pre-training (ResNet50+MGN) — 84.3 (2022-03-30) TAT — 82.5 (2022-04-01) BPBreID (RK) — 92.9 (2022-11-07) BPBreID — 84.2 (2022-11-07) CLIP-ReID (without re-ranking) — 83.1 (2022-11-25) DiP (without RK) — 85.2 (2022-12-24) MBNET — 83.58 (2023-02-28) PLIP-RN50-MGN — 81.7 (2023-05-15) CLIP-ReID Baseline+UFFM+AMC — 85.0 (2024-05-02) ReMix — 79.8 (2024-10-29) LOMO + XQDA — 17.04 (2014-06-17) OIM — 47.4 (2016-04-07) DLCE — 49.3 (2016-11-17) SVDNet — 56.8 (2017-03-16) PAN + re-rank — 66.74 (2017-07-03) PCB (RPP) — 69.2 (2017-11-26) PSE+ ECN (rank-dist) — 79.8 (2017-11-28) DaRe(De)+RE+RR [wang2018resource] — 80.0 (2018-05-22) SSP-ReID (RR) — 83.7 (2018-07-15) Parameter-Free Spatial Attention — 85.9 (2018-11-29) st-ReID(RE, RK,Cam) — 92.7 (2018-12-08) Unsupervised Pre-training (ResNet101+RK) — 92.77 (2020-12-07) DenseIL — 97.1 (2021-03-16)
RankModel mAPRank-1Rank-5Rank-10 Rank-1 Rank-5 Extra Training Data PaperCodeYear
1 DenseIL 97.1 Dense Interaction Learning for Video-based Person Re-identification 2021
2 CTL Model (ResNet50, 256x128) 96.195.696.297.9 On the Unreasonable Effectiveness of Centroids in Image Retrieval mikwieczorek/centroids-reid · lannguyen0910/deep-efficient-reid · lannguyen0910/deep-efficient-person-reid 2021
3 BPBreID (RK) 92.993.9 Body Part-Based Representation Learning for Occluded Person Re-Identification vlsomers/bpbreid · trackinglaboratory/tracklab · vlsomers/keypoint_promptable_reidentification 2022
4 Unsupervised Pre-training (ResNet101+RK) 92.7793.99 Unsupervised Pre-training for Person Re-identification DengpanFu/LUPerson 2020
5 st-ReID(RE, RK,Cam) 92.794.5 Spatial-Temporal Person Re-identification Wanggcong/Spatial-Temporal-Re-identification · SurajDonthi/Multi-Camera-Person-Re-Identification · BonaventureR/person-reid 2018
5 RGT&RGPR (RK) 92.794.3 Eliminate Deviation with Deviation for Data Augmentation and a General Multi-modal Data Learning Method finger-monkey/Data-Augmentation 2021
7 Viewpoint-Aware Loss(RK) 91.893.996.5 Viewpoint-Aware Loss with Angular Regularization for Person Re-Identification zzhsysu/VA-ReID 2019
8 LDS (ResNet50 + RK) 91.092.91 Learning to Disentangle Scenes for Person Re-identification deropty/LDS 2021
9 FlipReID (with re-ranking) 90.793.0 FlipReID: Closing the Gap between Training and Inference in Person Re-Identification nixingyang/FlipReID 2021
9 Adaptive L2 Regularization (with re-ranking) 90.792.2 Adaptive L2 Regularization in Person Re-Identification nixingyang/AdaptiveL2Regularization 2020
1–10 / 94 다음 → 페이지당 10 20 50 100