• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of Imaging Science and Technology >Random Occlusion Recovery for Person Re-identification
【24h】

Random Occlusion Recovery for Person Re-identification

机译：人重新识别的随机闭塞恢复

获取原文
获取原文并翻译 | 示例
       
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

As a basic task of multi-camera surveillance system, person re-identification aims to re-identify a query pedestrian observed from non-overlapping multiple cameras or across different time with a single camera. Recently, deep learning-based person re-identification models have achieved great success in many benchmarks. However, these supervised models require a large amount of labeled image data, and the process of manual labeling spends much manpower and time. In this study, we introduce a method to automatically synthesize labeled person images and adopt them to increase the sample number per identity for person re-identification datasets. To be specific, we use block rectangles to randomly occlude pedestrian images. Then, a generative adversarial network (GAN) model is proposed to use paired occluded and original images to synthesize the de-occluded images that are similar but not identical to the original image. Afterward, we annotate the de-occluded images with the same labels of their corresponding raw images and use them to augment the number of samples per identity. Finally, we use the augmented datasets to train baseline model. The experimental results on CUHK03, Market-1501 and DukeMTMC-relD datasets show the effectiveness of the proposed method. (C) 2019 Society for Imaging Science and Technology.
机译：作为多摄像机监控系统的基本任务，人重新识别旨在重新识别从非重叠多个摄像机或不同时间观察到的查询行人，或者用单个相机。最近，基于深度学习的人重新识别模型在许多基准中取得了巨大的成功。但是，这些监督模型需要大量标记的图像数据，手动标签的过程花费了许多人力和时间。在这项研究中，我们介绍了一种自动综合标记的人物图像的方法，并采用它们来增加人员重新识别数据集的每个身份的样本号。具体而言，我们使用块矩形来随机遮挡行人图像。然后，提出了一种生成的对抗性网络（GaN）模型来使用成对的遮挡和原始图像来合成类似但与原始图像相同而不是相同的去遮挡图像。之后，我们用相应的原始图像的相同标签向解闭合图像注释，并使用它们来增加每个身份的样本数量。最后，我们使用增强的数据集来训练基线模型。 CUHK03，市场-1501和DukemTMC-Reld数据集的实验结果表明了该方法的有效性。 （c）2019年成像科技协会。

著录项

  • 来源
    《Journal of Imaging Science and Technology》 |2019年第3期|030405.1-030405.9|共9页
  • 作者

    Wu Di; Zhang Kun; Zheng Si-Jia; Hao Yong-Tao; Liu Fu-Qiang; Qin Xiao; Cheng Fei; Zhao Yang; Liu Qi; Yuan Chang-An; Huang De-Shuang;

  • 作者单位

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

    Guangxi Teachers Educ Univ Sci Comp & Intelligent Informat Proc Guang Xi Hig Nanning 530001 Peoples R China;

    Beijing E Hualu Info Technol Co Ltd Beijing Peoples R China;

    Beijing E Hualu Info Technol Co Ltd Beijing Peoples R China;

    Edinburgh Napier Univ Sch Comp 10 Colinton Rd Edinburgh EH10 5IYI Midlothian Scotland;

    Guangxi Teachers Educ Univ Sci Comp & Intelligent Informat Proc Guang Xi Hig Nanning 530001 Peoples R China;

    Tongji Univ Sch Elect & Informat Engn Inst Machine Learning & Syst Biol Shanghai Peoples R China;

  • 收录信息
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号