
摘要
在过去的几年中,计算机视觉领域经历了一场主要由大规模数据集的出现和深度卷积神经网络用于端到端学习所推动的革命。人员再识别子领域也不例外。不幸的是,社区中普遍存在一种观点,即三元组损失(triplet loss)不如使用代理损失(分类、验证)后再进行单独的度量学习步骤。我们证明了,无论是从头开始训练的模型还是预训练的模型,使用三元组损失的一种变体进行端到端的深度度量学习都能大幅超越大多数其他已发表的方法。
代码仓库
agongt408/vbranch
tf
GitHub 中提及
jjmachan/DeepHash
pytorch
GitHub 中提及
BarbaAlGhul/recommender-system-thesis
GitHub 中提及
adambielski/siamese-triplet
pytorch
GitHub 中提及
bastiennNB/Pair_ReID
pytorch
GitHub 中提及
immuno121/audio_source_classification
pytorch
GitHub 中提及
cftang0827/pedestrian_recognition
tf
GitHub 中提及
kilsenp/triplet-reid-pytorch
pytorch
GitHub 中提及
zhengziqiang/ReshapeGAN
tf
GitHub 中提及
AdrianUng/keras-triplet-loss-mnist
tf
GitHub 中提及
thomas-liao/diva_tracking_reid
tf
GitHub 中提及
huanghoujing/AlignedReID-Re-Production-Pytorch
pytorch
GitHub 中提及
tbmoon/facenet
pytorch
GitHub 中提及
AsuradaYuci/tripletreid-zhushi
tf
GitHub 中提及
IcewineChen/mxnet-batch_hard_triplet_loss
tf
GitHub 中提及
cftang0827/human_recognition
tf
GitHub 中提及
shubhamtyagii/Aligned_Reid
pytorch
GitHub 中提及
VisualComputingInstitute/triplet-reid
tf
GitHub 中提及
letthedataconfess/Keras-triplet-loss-mnist
tf
GitHub 中提及
huynhtuan17ti/FaceNet-OneShotLearning
pytorch
GitHub 中提及
lyakaap/NetVLAD-pytorch
pytorch
GitHub 中提及
keetsky/Net_ghostVLAD-pytorch
pytorch
GitHub 中提及
layumi/Person_reID_baseline_pytorch
pytorch
GitHub 中提及
khrlimam/facenet
pytorch
GitHub 中提及
h4veFunCodin9/Aligned_ReID
pytorch
GitHub 中提及
abhishirk/Aligned_ReId
pytorch
GitHub 中提及
huanghoujing/person-reid-triplet-loss-baseline
pytorch
GitHub 中提及
BonaventureR/person-reid
pytorch
GitHub 中提及
基准测试
| 基准 | 方法 | 指标 |
|---|---|---|
| person-re-identification-on-cuhk03 | TriNet | Rank-1: 89.63 Rank-5: 99.01 |
| person-re-identification-on-dukemtmc-reid | TriNet | Rank-1: 72.44 mAP: 53.50 |
| person-re-identification-on-market-1501 | LuNet | Rank-1: 81.38 Rank-5: 92.34 mAP: 60.71 |
| person-re-identification-on-market-1501 | TriNet (RK) | Rank-1: 86.67 Rank-5: 93.38 mAP: 81.07 |
| person-re-identification-on-market-1501 | LuNet (RK) | Rank-1: 84.59 Rank-5: 91.89 mAP: 75.62 |
| person-re-identification-on-market-1501 | TriNet | Rank-1: 84.92 Rank-5: 94.21 mAP: 69.14 |
| person-re-identification-on-mars | TriNet | Rank-1: 79.80 Rank-5: 91.36 mAP: 67.70 |
| person-re-identification-on-mars | LuNet (RK) | Rank-1: 78.48 Rank-5: 88.74 mAP: 73.68 |
| person-re-identification-on-mars | TriNet (RK) | Rank-1: 81.21 Rank-5: 90.76 mAP: 77.43 |
| person-re-identification-on-mars | LuNet | Rank-1: 75.56 Rank-5: 89.70 mAP: 60.48 |