Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (7): 194-198.DOI: 10.3778/j.issn.1002-8331.1712-0256

Previous Articles     Next Articles

Pedestrian Detection Method Based on Ensemble SVM Classifier

GAN Ling1, YANG Meng2   

  1. 1.School of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
    2.School of Software Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • Online:2019-04-01 Published:2019-04-15

聚合支持向量机分类器的行人检测方法

甘  玲1,杨  梦2   

  1. 1.重庆邮电大学 计算机科学与技术学院,重庆 400065
    2.重庆邮电大学 软件工程学院,重庆 400065

Abstract: Since the accuracy is low due to the imbalance of positive and negative pedestrians in SVM pedestrian detection method which adopt under-sampling method, this paper proposes an EnsembleSVM pedestrian detection method by combining under-sampling and EasyEnsemble. Firstly, selecting negative sample as initial training sample randomly that is divided into multiple sub-negative sample sets equalizing the positive sample set, building balanced sub-training sets and linear assembling EasyEnsemble SVM. Then negative sample is classified and judged by using EasyEnsemble SVM, making misjudged sample as hard sample, building balanced sub-training sets again and training sub-classifier, combining which with EasyEnsemble SVM to get the Ensemble SVM classifier. Experiments on INRIA pedestrian data show the approach is better than classical SVM pedestrian detection algorithm in both detection speed and accuracy rate.

Key words: pedestrian detection, Support Vector Machine(SVM), EasyEnsemble SVM classifier, Ensemble SVM

摘要: 针对支持向量机分类器的行人检测方法采用欠采样方法,存在正负行人比例不平衡造成的准确率不高问题,结合欠采样和EasyEnsemble方法,提出一种聚合支持向量机(Ensemble SVM)分类器的行人检测方法。随机选择负样本作为初始训练样本,并将其划分为与正样本集均衡的多个子负样本集,构建平衡子训练集,线性组合成EasyEnsemble SVM分类器;利用该分类器对负样本进行分类判断,将误判样本作为难例样本,重新划分构建新的平衡子训练集,训练子分类器,结合EasyEnsemble SVM分类器,得到Ensemble SVM分类器行人检测方法。在INRIA行人数据集上的实验表明,该方法在检测速度和检测率上都优于经典的SVM行人检测算法。

关键词: 行人检测, 支持向量机(SVM), EasyEnsemble SVM分类器, 聚合支持向量机(Ensemble SVM)