《计算机应用研究》|Application Research of Computers

基于样本正态性重采样的改进KISSME行人再识别算法

Improved KISSME method for person re-identification based on normality resampling

免费全文下载 (已被下载 次)  
获取PDF全文
作者 宋丽丽,李彬,赵俊雅,刘国峰
机构 1.成都理工大学 工程技术学院,四川 乐山 614000;2.武汉轻工大学 机械工程学院,武汉 430023;3.武汉理工大学 理学院,武汉 430070
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)07-064-2227-05
DOI 10.19734/j.issn.1001-3695.2019.02.0098
摘要 跨场景行人再识别方法的关键在于特征识别和度量模型的建立,而这两方面的问题都受到图像样本分布的局限,进而使得模型参数的估计出现过拟合现象。针对以上跨场景的行人再识别问题,提出了一种基于半监督的改进KISSME算法。该算法在KISSME学习算法的基础上,根据样本数据的正态分布特性进行重采样,并通过构建循环优化的学习方式弱化模型的拟合强度,增强度量模型的泛化能力,以此建立泛化后的度量模型。再通过联合KISSME度量,构建改进的半监督度量模型。最后,利用行人再识别通用公开数据集VIPeR对改进算法的有效性进行验证,并与SLDDL、RDC、ITML、PCCA、QARR-RSVM和KISSME等算法精度相比较,实验结果表明基于半监督的改进KISSME算法在不同排名下都有明显的优势,尤其在rank-1识别精度上,相较于现有的KISSME算法提升了3.14%,充分验证了该算法的有效性。
关键词 行人再识别; 度量学习算法; 半监督学习
基金项目
本文URL http://www.arocmag.com/article/01-2020-07-064.html
英文标题 Improved KISSME method for person re-identification based on normality resampling
作者英文名 Song Lili, Li Bin, Zhao Junya, Liu Guofeng
机构英文名 1.College of Engineering & Technical,Chengdu University of Technology,Leshan Sichuan 614000,China;2.School of Mechanical Engineering,Wuhan Polytechnic University,Wuhan 430023,China;3.School of Science,Wuhan University of Technology,Wuhan 430070,China
英文摘要 As two critically important parts of the cross-camera pedestrian re-recognition method, feature recognition and metric model establishment have been constrained by notorious overfitting of model parameter estimation that arises from improper image sample distribution. This study proposed an improved, semi-supervised KISSME learning algorithm-based method for pedestrian re-recognition. The proposed method succeeded to construct a generalized measurement model by re-sampling normally distributed data, weakening fitting strength through establishment of a circular optimization metric learning method, and improving the model's generalization capacity. Then, it introduced KISSME metrics to further improve the semi-supervised model. Finally, it verified effectiveness of the improved algorithm by pedestrian re-recognition using the public open VIPeR dataset, results of which were compared with accuracies of SLDDL, RDC, ITML, PCCA, QARR-RSVM and KISSME. It demonstrated the improved, semi-supervised KISSME algorithm to be superior in all recognition accuracy ranks, especially in the rank-1. It achieved an accuracy that was 3.14% higher than that of the existing KISSME algorithm, thereby validating effectiveness of the algorithm proposed in this study.
英文关键词 person re-identification; metric learning; semi-supervised
参考文献 查看稿件参考文献
 
收稿日期 2019/2/28
修回日期 2019/4/18
页码 2227-2231
中图分类号 TP391.9
文献标志码 A