| 加权极限学习机在行人检测中的研究和应用 |
| |
| 引用本文： | 刘倩,李策,杨峰,刘立波,邓箴.加权极限学习机在行人检测中的研究和应用[J].计算机工程与设计,2019,40(8):2366-2371. |
| |
| 作者姓名： | 刘倩 李策 杨峰 刘立波 邓箴 |
| |
| 作者单位： | 中国矿业大学(北京)机电与信息工程学院,北京100083;宁夏大学信息工程学院,宁夏银川750001;中国矿业大学(北京)机电与信息工程学院,北京,100083;宁夏大学信息工程学院,宁夏银川,750001 |
| |
| 基金项目： | 贵州省科技计划;国家自然科学基金;宁夏回族自治区重点研发计划(重点)基金项目 |
| |
| 摘 要： | 针对AdaBoost框架下加权极限学习机没有考虑到离群点对分类面的影响,提出基于AdaBoost的权值极限学习机的改进算法。利用隐层空间样本间欧式距离设置学习机初始权值,根据信息熵相似性原理更新AdaBoost框架内弱分类器权值。在INRIA和Caltech-USA两个不平衡行人数据库上的分析实验结果表明,该算法具有原算法处理不平衡样本分类的能力,同时误检率得到显著降低。
|
| 关 键 词： | 加权极限学习机 自适应增强 信息熵 相似性原理 行人检测 |
Research and application of weighted extreme learning machine in pedestrian detection |
| |
| LIU Qian,LI Ce,YANG Feng,LIU Li-bo,DENG Zhen.Research and application of weighted extreme learning machine in pedestrian detection[J].Computer Engineering and Design,2019,40(8):2366-2371. |
| |
| Authors: | LIU Qian LI Ce YANG Feng LIU Li-bo DENG Zhen |
| |
| Affiliation: | (School of Mechanical Electronic and Information Engineering,China University of Mining and Technology,Beijing 100083,China;College of Information Engineering,Ningxia University,Yinchuan 750001,China) |
| |
| Abstract: | Boosting weighted extreme learning machine,in which weighted extreme learning machine is embedded into AdaBoost framework,reckons without the influence of outliers.To circumvent these problems for previous algorithm,an improved approach based on boosting weighted extreme learning machine was proposed,for rapid detection of pedestrian.The initial value of weight was set according to Euclidean distance between samples mapped into hidden layer space,and comparability measure based on information entropy was utilized to update the weight of weak classifiers in AdaBoost framework.Experimental results over the INRIA and Caltech-USA pedestrian datasets show that the proposed approach has the ability of generalization like pre- vious algorithm,and reduces the detection error rate significantly. |
| |
| Keywords: | weighted extreme learning machine AdaBoost information entropy similarity principle pedestrian detection |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
