Context feature fusion and enhanced non-maximum suppression for pedestrian detection in crowded scenes
文献类型：期刊论文
| 作者 | Shao, Yu2; Hu, Jianhua1 ; Hu, Lihua2; Zhang, Jifu2; Wang, Xinbo1
|
| 刊名 | MULTIMEDIA TOOLS AND APPLICATIONS
 |
| 出版日期 | 2024-03-16 |
| 页码 | 21 |
| 关键词 | Densely populated Pedestrian detection Occlusion Contextual information |
| ISSN号 | 1380-7501 |
| DOI | 10.1007/s11042-024-18865-x |
| 通讯作者 | Hu, Lihua(hlh@tyust.edu.cn) |
| 英文摘要 | Pedestrian detection has a wide range of applications in the field of multimedia, and significant progress has been made. However, in densely populated scenes, there are two problems: occlusion and mistake suppression of overlapping bounding boxes, which lead to false positives and false negatives, thereby degrading overall performance. To tackle these problems, firstly, by leveraging contextual information to capture correlations between pedestrians and backgrounds, we propose the Context Feature Fusion Module (CFFM), which alleviates the absence of crucial features caused by occlusion. Secondly, by combining the intersection over Union (IoU) and the distance between center points of overlapping bounding boxes, we propose Distance Set Non-Maximization Suppression (DSNMS), which tackles error suppression of overlapping bounding boxes. Finally, extensive experiments were conducted on the CrowdHuman dataset, yielding remarkable results for our method with an Average Precision (AP) of 91.22%, a Log average miss rate (MR-2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$<^>{-2}$$\end{document}) of 40.26%, and a Jaccard Index (JI) of 83.54%. Furthermore, the visualization results of real-world scenes further validate the efficacy of our proposed method. |
| WOS关键词 | NMS |
| 资助项目 | National Natural Science Foundation of China[62273248] ; National Natural Science Foundation of China (NSFC)[JD2022005] ; Computer Vision Joint Training Demonstration Base of Taiyuan University of Science and Technology |
| WOS研究方向 | Computer Science ; Engineering |
| 语种 | 英语 |
| WOS记录号 | WOS:001184608000005 |
| 出版者 | SPRINGER |
| 资助机构 | National Natural Science Foundation of China ; National Natural Science Foundation of China (NSFC) ; Computer Vision Joint Training Demonstration Base of Taiyuan University of Science and Technology |
| 源URL | [http://ir.ia.ac.cn/handle/173211/56960]  |
| 专题 | 智能制造技术与系统研究中心_先进制造与自动化 |
| 通讯作者 | Hu, Lihua |
| 作者单位 | 1.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China 2.Taiyuan Univ Sci & Technol, Sch Comp Sci & Technol, Waliu Rd, Taiyuan 030024, Shanxi, Peoples R China |
| 推荐引用方式 GB/T 7714 | Shao, Yu,Hu, Jianhua,Hu, Lihua,et al. Context feature fusion and enhanced non-maximum suppression for pedestrian detection in crowded scenes[J]. MULTIMEDIA TOOLS AND APPLICATIONS,2024:21. |
| APA | Shao, Yu,Hu, Jianhua,Hu, Lihua,Zhang, Jifu,&Wang, Xinbo.(2024).Context feature fusion and enhanced non-maximum suppression for pedestrian detection in crowded scenes.MULTIMEDIA TOOLS AND APPLICATIONS,21. |
| MLA | Shao, Yu,et al."Context feature fusion and enhanced non-maximum suppression for pedestrian detection in crowded scenes".MULTIMEDIA TOOLS AND APPLICATIONS (2024):21. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。