People Re-Identification with Local Distance Comparison Using Learned Metric
People Re-Identification with Local Distance Comparison Using Learned Metric
复制标题
DOI：
10.1587/transinf.2013edp7424
复制
发表时间：
2014-09
期刊：
影响因子：
--
通讯作者：
Guanwen Zhang;Jien Kato;Yu Wang;K. Mase
中科院分区：
文献类型：
--
作者：
Guanwen Zhang;Jien Kato;Yu Wang;K. Mase
在本文中，我们提出了一种新的方法，多镜头的人重新识别。由于相机视图、光照、姿态的非刚性变形等的高方差，存在关键的互方差/内方差问题，即，同一个人看起来可能差别很大，而不同的人看起来可能极其相似。这个问题导致了一个棘手的，多模态分布的人的外观特征空间。为了处理数据的多模态特性，我们在局部距离比较框架下解决了重新识别问题，这大大减少了每个人的不同外观所引起的困难。此外，我们建立了一个基于能量的损失函数，通过计算特征空间中相应子集之间的距离来衡量外观实例之间的相似性。该损失函数不仅有利于指示相同人的外观之间的高相似性的小距离，而且还惩罚反映不同人的外观之间的高相似性的子集之间的小距离或不期望的重叠。以这种方式，有效的人员重新识别可以针对差异间/差异内问题以鲁棒的方式实现。我们的方法的性能进行了评估，将其应用到公共基准数据集ETHZ和CAVIAR 4 REID。实验结果表明，显着的改善，比以前的报告。关键词：多炮重识别，局部距离比较，多峰分布
In this paper, we propose a novel approach for multipleshot people re-identification. Due to high variance in camera view, light illumination, non-rigid deformation of posture and so on, there exists a crucial inter-/intravariance issue, i.e., the same people may look considerably different, whereas different people may look extremely similar. This issue leads to an intractable, multimodal distribution of people appearance in feature space. To deal with such multimodal properties of data, we solve the re-identification problem under a local distance comparison framework, which significantly alleviates the difficulty induced by varying appearance of each individual. Furthermore, we build an energy-based loss function to measure the similarity between appearance instances, by calculating the distance between corresponding subsets in feature space. This loss function not only favors small distances that indicate high similarity between appearances of the same people, but also penalizes small distances or undesirable overlaps between subsets, which reflect high similarity between appearances of different people. In this way, effective people re-identification can be achieved in a robust manner against the inter-/intravariance issue. The performance of our approach has been evaluated by applying it to the public benchmark datasets ETHZ and CAVIAR4REID. Experimental results show significant improvements over previous reports. key words: multiple-shot re-identification, local distance comparison, multimodal distribution