一种改进的增强型AdaBoost算法
首发时间:2010-01-19
摘要:本文分析了传统的Adaboost算法在训练过程中可能出现的退化问题以及目标类权重分布出现过适应的现象,文章提出了一种改进的Adaboost算法,该算法基于调整加权误差分布,限制权重扩张。另外,改进了最终强分类器输出形式,实验结果表明,改进后的Adaboost算法在测试视频集中大幅度提高了检测率、降低了误检率.在家庭视频安防监控领域具有较好的应用前景。
For information in English, please click here
An Improved Enhanced AdaBoost Algorithm
Abstract:This paper analyzes the degradation issues as well as the distribution of target class weights appeared to adapt to the phenomenon that may arise in the training process of the traditional Adaboost algorithm. The article proposes a modified Adaboost algorithm based on the adjusted weighted error distribution, limiting the expansion weights,and improves the final strong classifier output forms.Experiment shows that the improved Adaboost algorithm increased the detection rate and decreased the false positive rate when we test the video set.There are good prospects of application in the field of home video security surveillance.
Keywords: Adaboost Weight distribution Weighted error Strong classifier
基金:
论文图表:
引用
No.3916551312112638****
同行评议
共计0人参与
勘误表
一种改进的增强型AdaBoost算法
评论
全部评论0/1000