《计算机应用研究》|Application Research of Computers

基于似物目标的快速行人检测算法

Rapid pedestrian detection based on generic object generation

免费全文下载 (已被下载 次)  
获取PDF全文
作者 刘倩,李策,杨峰,刘立波
机构 1.中国矿业大学(北京)机电与信息工程学院,北京 100083;2.宁夏大学 信息工程学院,银川 750001
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2019)07-063-2219-04
DOI 10.19734/j.issn.1001-3695.2018.01.0115
摘要 针对传统滑动窗口行人检测速度慢、无效窗口数量大和识别率低的问题,提出基于似物目标的快速行人检测算法。该算法首先使用改进的二值化赋范梯度算法生成目标候选集,再利用NDOG特征描述子计算目标特征。通过在INRIA和Caltech-USA两个行人数据库的分析实验,证明该算法可有效减少待检测窗口数量,提高特征表示能力,且计算时间显著降低,使用线性支持向量机作为分类器可获得较高的行人检测率。
关键词 似物目标检测; 二值化赋范梯度; 方向梯度直方图; 极限学习机
基金项目 贵州省科技计划资助项目(黔科合GZ字[2015]3020)
国家自然科学基金资助项目(61751215)
本文URL http://www.arocmag.com/article/01-2019-07-063.html
英文标题 Rapid pedestrian detection based on generic object generation
作者英文名 Liu Qian, Li Ce, Yang Feng, Liu Libo
机构英文名 1.School of Mechanical Electronic & Information Engineering,China University of Mining & Technology,Beijing 100083,China;2.College of Information Engineering,Ningxia University,Yinchuan 750001,China
英文摘要 Detection of pedestrians on the image using a sliding window mode results in a huge number of invalid windows and inefficiency. To circumvent these problems for pedestrian detection, this paper proposed an approach, based on generic object generation, for rapid detection of pedestrian. Firstly, the approach generated the generic object proposals by the improved binarized normed gradients, and then calculated the character of proposals based on this NDOG feature descriptor. Experimental results over the INRIA and Caltech-USA pedestrian datasets show that this approach reduces the number of proposals effectively, and improves the efficiency and the detection accuracy significantly along with a linear SVM classifier.
英文关键词 generic object proposal generation; histogram of oriented gradient(HOG); binarized normed gradients(BING); extreme learnng machine(ELM)
参考文献 查看稿件参考文献
 
收稿日期 2018/1/20
修回日期 2018/3/9
页码 2219-2222
中图分类号 TP391.41
文献标志码 A