Mining Hard Samples Globally and Efficiently for Person Reidentification
Mining Hard Samples Globally and Efficiently for Person Reidentification
复制标题
在全球范围内高效地挖掘硬样本以进行人员重新识别
DOI：
10.1109/jiot.2020.2980549
复制
发表时间：
2020-10
影响因子：
10.6
通讯作者：
Zhang Xiong
中科院分区：
文献类型：
--
作者：
Hao Sheng;Yanwei Zheng;Wei Ke;Dongxiao Yu;Xiuzhen Cheng;Weifeng Lyu;Zhang Xiong
人员身份识别（ReID）是物联网（IoT）的一个重要应用。ReID在不同位置和时间识别摄像头视图中的行人，这通常被视为一项排名任务。这个任务的一个重要部分是硬样本挖掘。从技术上讲，可以采用两种策略，即，全局硬采矿和局部硬采矿。对于前者，在整个训练集中挖掘硬样本，而对于后者，它是在小批量中完成的。在文献中，大多数现有的方法在本地操作。例子包括批量硬样本挖掘和半硬样本挖掘。很少使用全局硬挖掘的原因是计算复杂度高。在本文中，我们认为全局挖掘有助于找到更难的样本，从而有利于模型训练。为此，本文介绍了一种新的系统：1）从整个训练集中有效地挖掘硬样本（阳性和阴性），2）在训练中有效地使用它们。具体地说，提出了一种与多重态损失耦合的排序列表网络。一方面，多重态损失使得排序列表逐步创建，避免了耗时的初始化。另一方面，多重态损失的目的是在训练过程中有效地利用硬样本和易样本。此外，排名列表可以在全球范围内有效地挖掘硬性阳性和阴性样本。在实验中，我们探索了全局和局部样本挖掘方法的性能，以及半硬，最硬和随机选择的样本的影响。最后，我们使用各种公共数据集证明了我们的理论的有效性，并通过定量评估取得了有竞争力的结果。
Person reidentification (ReID) is an important application of Internet of Things (IoT). ReID recognizes pedestrians across camera views at different locations and time, which is usually treated as a ranking task. An essential part of this task is the hard sample mining. Technically, two strategies could be employed, i.e., global hard mining and local hard mining. For the former, hard samples are mined within the entire training set, while for the latter, it is done in mini-batches. In literature, most existing methods operate locally. Examples include batch-hard sample mining and semihard sample mining. The reason for the rare use of global hard mining is the high computational complexity. In this article, we argue that global mining helps to find harder samples that benefit model training. To this end, this article introduces a new system to: 1) efficiently mine hard samples (positive and negative) from the entire training set and 2) effectively use them in training. Specifically, a ranking list network coupled with a multiplet loss is proposed. On the one hand, the multiplet loss makes the ranking list progressively created to avoid the time-consuming initialization. On the other hand, the multiplet loss aims to make effective use of the hard and easy samples during training. In addition, the ranking list makes it possible to globally and effectively mine hard positive and negative samples. In the experiments, we explore the performance of the global and local sample mining methods, and the effects of the semihard, the hardest, and the randomly selected samples. Finally, we demonstrate the validity of our theories using various public data sets and achieve competitive results via a quantitative evaluation.
登录
查看更多内容
DOI：
10.1109/cvprw.2017.186
发表时间：
2017-07
期刊：
2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
影响因子：
--
作者：
Arne Schumann;R. Stiefelhagen
通讯作者：
Arne Schumann;R. Stiefelhagen
DOI：
10.1109/cvpr.2017.566
发表时间：
2017-07
期刊：
2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子：
--
作者：
Jiaxin Chen;Yunhong Wang;Jie Qin;Li Liu;Ling Shao
通讯作者：
Jiaxin Chen;Yunhong Wang;Jie Qin;Li Liu;Ling Shao
影响因子：
6.6
作者：
Huan Yang;Feng Li;Dongxiao Yu;Yifei Zou;Jiguo Yu
通讯作者：
Huan Yang;Feng Li;Dongxiao Yu;Yifei Zou;Jiguo Yu
DOI：
--
发表时间：
2017-03
期刊：
ArXiv
影响因子：
--
作者：
Alexander Hermans;Lucas Beyer;B. Leibe
通讯作者：
Alexander Hermans;Lucas Beyer;B. Leibe
DOI：
10.1109/iccv.2017.550
发表时间：
2017-09
期刊：
2017 IEEE International Conference on Computer Vision (ICCV)
影响因子：
--
作者：
Mang Ye;A. J. Ma;Liang Zheng;Jiawei Li;P. Yuen
通讯作者：
Mang Ye;A. J. Ma;Liang Zheng;Jiawei Li;P. Yuen