›› 2015, Vol. 28 ›› Issue (12): 80-.

• 论文 • 上一篇    下一篇

基于肤色模型和改进Adaboost算法的人脸检测

赵喆,侯俊   

  1. (上海理工大学 光电信息与计算机工程学院,上海 200093)
  • 出版日期:2015-12-15 发布日期:2015-12-15
  • 作者简介:赵喆(1991—),男,硕士研究生。研究方向:图像处理与识别。E-mail:334090445@qq.com。侯俊(1971—),女,博士。研究方向:图像处理,信号处理。

Face Detection Based on YCbCr Skin Color Model and Improved Adaboost Algorithm

ZHAO Zhe,HOU Jun   

  1. (School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)
  • Online:2015-12-15 Published:2015-12-15

摘要:

为提高人脸检测的精度,提出一种基于YCbCr肤色模型和Adaboost算法的人脸检测方法。利用YCbCr颜色空间的肤色阈值模型确定待检测区域。再利用Adaboost算法准确定位出人脸位置,并对Adaboost算法的训练过程提出一种新的权值更新方法,防止权值过分增大,避免了Adaboost算法训练过程中出现的“退化”现象。

关键词: YCbCr肤色模型, Adaboost算法, 人脸检测, 权值更新抑制

Abstract:

This paper proposes a method based on YCbCr color model and Adaboost face detection to improve the accuracy of face detection.Firstly,the algorithm uses the YCbCr skin color threshold model to determine the area to be detected.Then Adaboost algorithm is applied to locate the position of face.To avoid the degradation during training in Adaboost,the paper proposes a new method about suppression of weights updating in training of Adaboost algorithm.

Key words: YCbCr skin color model;Adaboost;face detection;suppression of weights updating