计算机工程 ›› 2020, Vol. 46 ›› Issue (5): 247-253.doi: 10.19678/j.issn.1000-3428.0056642

• 图形图像处理 • 上一篇    下一篇

基于姿态引导对齐网络的局部行人再识别

郑烨, 赵杰煜, 王翀, 张毅   

  1. 宁波大学 信息科学与工程学院, 浙江 宁波 315000
  • 收稿日期:2019-11-19 修回日期:2020-01-13 发布日期:2020-01-16
  • 作者简介:郑烨(1994-),女,硕士研究生,主研方向为计算机视觉;赵杰煜,教授;王翀,副教授;张毅,硕士研究生。
  • 基金项目:
    国家自然科学基金(61603202,61571247);浙江省自然科学基金重点项目(LZ16F03001,LY17F030002)。

Partial Pedestrain Re-Identification Based on Pose-Guided Alignment Network

ZHENG Ye, ZHAO Jieyu, WANG Chong, ZHANG Yi   

  1. College of Information Science and Engineering, Ningbo University, Ningbo, Zhejiang 315000, China
  • Received:2019-11-19 Revised:2020-01-13 Published:2020-01-16

摘要: 将局部行人再识别中的局部图像与整体图像直接进行比较会产生严重的空间错位,从而导致无法检测到正确目标。针对相同尺寸的行人局部图像与全局图像不匹配问题,提出姿态引导对齐网络(PGAN)模型,将姿态作为辅助信息引入到姿态引导的空间变换模块中,从局部图像与整体图像中提取仿射变换后的行人图像并将其与标准姿态进行对齐,再利用卷积神经网络学习相关特征实现局部行人再识别。实验结果表明,在Partial-REID数据集上PGAN模型取得65%的Rank-1准确率,相比直接使用深度卷积神经网络提取全局特征进行匹配的基准模型提高了3.7%,从而证明其具有良好的局部图像对齐能力及行人再识别效果。

关键词: 局部行人再识别, 对齐网络, 空间变换, 姿态, 深度卷积神经网络

Abstract: In partial pedestrian re-identification,serious spatial misalignment will be caused when the partial image of a pedestrian is directly compared with the holistic image,leading to a failure in target detection.To solve the mismatch of the partial pedestrian image and the holistic image of the same size,this paper proposes a Pose-Guided Alignment Network(PGAN) model.The PGAN firstly introduces the pose into Pose-Guided Spatial Transformation(PST) module as auxiliary information,extracts the pedestrian image after affine transformation from the partial image and holistic image,and compares the pedestrian image with the standard pose.Then the Convolutional Neural Network(CNN) is used to learn the features for partial pedestrian re-identification.Experimental results on the Partial-REID dataset show that the rank-1 accuracy of the PGAN model reaches 65%,which is 3.7% higher than that of the baseline model that directly extracts the global features with Deep Convolutional Neural Network(DCNN).The results demonstrate the proposed model has excellent performance in partial image alignment and pedestrian re-identification.

Key words: partial pedestrian re-identification, alignment network, spatial transformation, pose, Deep Convolutional Neural Network(DCNN)

中图分类号: