基于不变特征和核距离度量学习的行人再识别.pdfVIP

Journal of Image and Signal Processing 图像与信号处理, 2018, 7(2), 65-73 Published Online April 2018 in Hans. /journal/jisp /10.12677/jisp.2018.72008 Invariant Feature and Kernel Distance Metric Learning Based Person Re-Identification 1,2 2 1 Qi Liu , Li Hou , Zhangyou Peng 1 School of Communication Information Engineering, Shanghai University, Shanghai 2 School of Information Engineering, Huangshan University, Huangshan Anhui th th th Received: Mar. 15 , 2018; accepted: Mar. 28 , 2018; published: Apr. 4 , 2018 Abstract Pedestrian may vary greatly in appearance due to differences in illumination, viewpoint, and pose across cameras, which can bring serious challenges in person re-identification. A kernel distance metric learning algorithm of invariant feature is proposed for person re-identification in this pa- per. Firstly, an invariant feature composed of a concatenation of local maximal occurrence (LOMO) and feature fusion net (FFN) called LOMO-FFN is used to encode human appearance across cam- eras. Secondly, a gauss kernel distance metric learning algorithm called kernel canonical correla- tion analysis (KCCA) is applied to obtain an optimized human feature distance across cameras, based on the extracted feature representation. Experimental results have shown that the pro- posed algorithm effectively improves recognition rates on two challenging datasets (VIPeR, PRID450s). Keywords Person Re-Identification, Invariant Feature, Kernel Distance Metric Learning 基于不变特征和核距离度量学习的行人再识别 1,2 2 1 刘 琦 ，侯 丽 ，彭章友 1上海大学通信与信息工程学院，上海 2黄山学院信息工程学院，安徽 黄山