当前位置: X-MOL 学术Entropy › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Threshold-Based Hierarchical Clustering for Person Re-Identification
Entropy ( IF 2.7 ) Pub Date : 2021-04-24 , DOI: 10.3390/e23050522
Minhui Hu , Kaiwei Zeng , Yaohua Wang , Yang Guo

Unsupervised domain adaptation is a challenging task in person re-identification (re-ID). Recently, cluster-based methods achieve good performance; clustering and training are two important phases in these methods. For clustering, one major issue of existing methods is that they do not fully exploit the information in outliers by either discarding outliers in clusters or simply merging outliers. For training, existing methods only use source features for pretraining and target features for fine-tuning and do not make full use of all valuable information in source datasets and target datasets. To solve these problems, we propose a Threshold-based Hierarchical clustering method with Contrastive loss (THC). There are two features of THC: (1) it regards outliers as single-sample clusters to participate in training. It well preserves the information in outliers without setting cluster number and combines advantages of existing clustering methods; (2) it uses contrastive loss to make full use of all valuable information, including source-class centroids, target-cluster centroids and single-sample clusters, thus achieving better performance. We conduct extensive experiments on Market-1501, DukeMTMC-reID and MSMT17. Results show our method achieves state of the art.

中文翻译:

基于阈值的层次聚类用于人员重新识别

在人员重新识别(re-ID)中,无监督域自适应是一项艰巨的任务。最近,基于集群的方法取得了良好的性能。聚类和训练是这些方法中的两个重要阶段。对于聚类而言,现有方法的一个主要问题是它们无法通过丢弃聚类中的离群值或仅合并离群值来充分利用离群值中的信息。对于训练,现有方法仅将源特征用于预训练,将目标特征用于微调,并且没有充分利用源数据集和目标数据集中的所有有价值的信息。为了解决这些问题,我们提出了Ť基于hreshold ħ与ierarchical聚类方法Ç交易损失(THC)。THC具有两个特征:(1)将异常值视为参与训练的单样本聚类。它可以很好地将异常信息保存在离群值中,而无需设置聚类数目,并且结合了现有聚类方法的优点;(2)它利用对比损失来充分利用所有有价值的信息,包括源类质心,目标聚类质心和单样本聚类,从而获得更好的性能。我们在Market-1501,DukeMTMC-reID和MSMT17上进行了广泛的实验。结果表明我们的方法达到了最先进的水平。
更新日期:2021-04-24
down
wechat
bug