摘要
提出了一种基于肤色和改进AdaBoost算法的人脸检测方法。从被检图像中分割出肤色区域从而得到包含一系列人脸静态特征的候选人脸区域。针对传统AdaBoost算法在训练过程中的过增益现象提出了一种新的权重更新方法,同时在训练过程中构建级联分类器。通过级联分类器对候选人脸区域进行扫描来准确定位人脸。大量的实验结果表明,所提出的方法在人脸检测上取得了较好的效果。
This paper proposes a face detection algorithm combined with skin color detection and improved AdaBoost algorithm.First,skin regions are segmented from the detected image,and candidate face regions are obtained in terms of the statistical characteristics of human face;Then focusing on the phenomena of overfitting in training process of classical AdaBoost algorithm,this paper proposes a novel method to update weight.At the same time,the process of constructing cascade classifier is added to training process.Finally,the candidate face regions are scanned by cascade classifier for more exact face orientation.A mass of experimental results show that the new approach obtains better results and improves detection performance obviously.
出处
《重庆理工大学学报(自然科学)》
CAS
2012年第11期93-97,共5页
Journal of Chongqing University of Technology:Natural Science