Pedestrian re-identification based on feature fusion and improved neural network

JIANG Hui-hui; ZHANG Rong; LI Xiao-bao; GUO Li-jun

doi:10.13873/J.1000-9787(2017)08-0121-05

Chinese

您当前的位置：

首页 >

文章列表页 >

Pedestrian re-identification based on feature fusion and improved neural network

更新时间：2025-09-11

- Pedestrian re-identification based on feature fusion and improved neural network
- Vol. 36, Issue 8, Pages: 121-125(2017)
- 作者机构：
  
  宁波大学信息科学与工程学院
- 作者简介：
- 基金信息：
- DOI：10.13873/J.1000-9787(2017)08-0121-05
  CLC： TP183;TP391.41
- Published：2017
- 稿件说明：
移动端阅览
[1]蒋桧慧,张荣,李小宝,郭立君.基于特征融合与改进神经网络的行人再识别[J].传感器与微系统,2017,36(08):121-125.

JIANG Hui-hui, ZHANG Rong, LI Xiao-bao, et al. Pedestrian re-identification based on feature fusion and improved neural network[J]. 2017, 36(8): 121-125.
[1]蒋桧慧,张荣,李小宝,郭立君.基于特征融合与改进神经网络的行人再识别[J].传感器与微系统,2017,36(08):121-125. DOI： 10.13873/J.1000-9787(2017)08-0121-05.

JIANG Hui-hui, ZHANG Rong, LI Xiao-bao, et al. Pedestrian re-identification based on feature fusion and improved neural network[J]. 2017, 36(8): 121-125. DOI： 10.13873/J.1000-9787(2017)08-0121-05.

摘要

行人再识别中

为了获得基于突出性颜色名称的颜色描述（SCNCD）特征对于光照变化较好的鲁棒性

提出了融合SCNCD特征和对于视角变化鲁棒性高的局部最大出现概率（LOMO）表观特征;为了获得图像的结构信息

将图像划分为多个重叠块

并提取块特征;针对神经网络容易陷入局部极小值

且收敛速度慢的问题

引入动量项。经过公用VIPeR数据库和PRID450s数据测试后

实验结果表明:融合后的特征的识别能力明显高于原特征的识别能力

且改进后的神经网络收敛速度明显提高。

Abstract

In person re-identification

aiming at salient color named based color descriptor（ SCNCD） features based on salient color names are robust to illumination changes

propose to fuse the SCNCD features and the LOMO features which are robust to viewpoint changes. In order to get structure information of images

the images are divided into overlapping patches and the patch features are extracted. In order to solve the problems that neural network is easy to fall into local minimum and its convergence speed is slow

momentum term is introduced. The proposed method has been tested in the most challenging public VIPeR database and PRID450 s database

and experimental results prove that recognition abilities of the fused features are obviously higher than that of the original features

and convergence speed of the improved neural network is increased obviously.

关键词

Keywords

references

基于距离匹配的行人再识别技术综述 [J]. 俞婧,仇春春,王恬,许金鑫. 微处理机 . 2016(03)

智能视频监控技术综述 [J]. 黄凯奇,陈晓棠,康运锋,谭铁牛. 计算机学报 . 2015(06)

Covariance descriptor based on bio-inspired features for person re-identification and face verification [J] . Bingpeng Ma,Yu Su,Frédéric Jurie. Image and Vision Computing . 2014 (6-7)

Views

202

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research progress of surface electromyography muscle strength estimation model

Research on gas sensor humidity compensation based on GA-BP neural network

Research on human action recognition and contrast based on OpenPose

Pedestrian re-identification based on scenes transition and region alignment

Improved sensor fault diagnosis and fault tolerance of neural network

Related Author

YU Fengfan

WEI Dejian

FENG Yanyan

MA Yifan

LI Zhenjiang

LIAO Youwei

CAO Weijia

JIANG Jianfeng

Related Institution

College of Intelligence and Information Engineering,Shandong University of Traditional Chinese Medicine

Key laboratory of Wireless Sensor Networks in Yunnan Higher Education Institutions

School of Electrical and Information Technology, Yunnan University of Nationalities

School of Information Science and Engineering, Wuhan University of Science and Technology

School of Electronic and Information Engineering,Hebei University of Technology

⁰