摘要
针对不同摄像机场景中的行人图像受到光照、视角和行人姿态等变化的影响,在监控画面中容易造成较大的外观差异的问题,提出了一种基于核空间与稠密水平条带特征的行人再识别算法;该算法在XQDA(Cross-view Quadratic Discriminant Analysis)度量学习算法的基础上提出了核空间映射与稠密水平条带提取行人图像特征的思想,首先通过自顶向下的滑动水平条带提取每个水平条带的颜色特征和纹理特征,然后融合行人图像的多种特征,把获得的特征映射到核空间中,最后在核空间里学习得到一个对背景、视角、姿势的变化具有鲁棒性的相似度函数,通过比较相似度的排名来对行人进行再识别;在VIPeR和iLIDS两个行人再识别数据集上的实验结果表明,该算法具有较高的识别率,其中Rank1(排名第1的搜索结果即为待查询行人的比率)分别达到48.2%和60.8%。
A new person re-identification algorithm based on dense horizontal stripes and kernel space mapping for several problems occurred in person images of different camera views,such as illumination changes,different viewpoints and varying poses,it is likely to form a lot of differences in appearance.The proposed method propose the idea of kernel space mapping and dense horizontal stripes extraction of person images based on the Cross-view Quadratic Discriminant Analysis(XQDA)metric learning algorithm.First,the each horizontal stripe of person images is extracted from color features and a texture feature by using the top-down sliding horizontal stripe.Then,fusing multi-features of person images and mapping the obtained features to kernel space.Finally,the proposed algorithm gets a similarity function which is robust to the change of background,viewpoint and posture by learning in kernel space.Pedestrians are re-identified by comparing rankings of similarities.The proposed method is demonstrated on two public benchmark datasets including VIPeR and iLIDS,and experimental results show that the proposed method achieves excellent re-identification rates compared with other similar algorithms.Moreover,the proposed method achieves a 48.2%at rank1(represents the correct matched pair)on VIPeR benchmark and a 60.8%at rank1 on iLIDS benchmark respectively.
作者
王强
包晓安
张福星
高春波
桂江生
Wang Qiang;Bao Xiaoan;Zhang Fuxing;Gao Chunbo;Gui Jiangsheng(School of Information,Zhejiang Sci-Tech University,Hangzhou 310018,China)
出处
《计算机测量与控制》
2018年第7期173-177,共5页
Computer Measurement &Control
基金
国家自然科学基金(61379036
61502430)
国家自然科学基金委中丹合作项目(61361136002)
浙江省重大科技专项重点工业项目(2014C01047)
浙江理工大学521人才培养计划(20150428)
关键词
行人再识别
距离度量学习
核空间
滑动水平条带
person re-identification
metric learning
kernel space
slide horizontal stripe
