一种基于Haar小波变换的彩色图像人脸检测方法

本文提出了一种基于Haar小波变换的彩色图像人脸检测方法。首先进行彩色空间的变换,检测出图像中的肤色区域;利用Adaboost算法训练出的基于Haar小波变换的检测器对该区域进行人脸检测;建立眼睛颜色模型,并根据眼睛!嘴在不同分量上的分布特征,将它们从人脸区域中提取出来;最后融合眼睛!嘴候选区域的信息,利用特征不变的方法进行眼睛和嘴的确定。实验结果表明,该方法能够快速有效地检测出人脸,并能够确定眼!嘴的位置。

An Approach of Face Detection In Color Images Based on Haar Wavelet

This paper presents an approach of face detection in color image based on Haar wavelet. First,color space transforma- tion is used to detect skin color regions, then we train a Haar- based face detector using Adaboost learning algorithm and con- struct an eye color model and extract eye and mouth candidates based on the distribution features on each component of color space. Finally we fuse the information of each candidate and validate faces using feature invariant method. Experimental re- sults show that this algorithm is effective in detecting faces as well as localizing mouth and eyes.