基于Haar特征和肤色特征的一种面部检测方法

An Improved Face Recognition Approach Based on a Combination of Haar Feature and Skin Color Detection

本文提出用于人脸识别系统(Face Recognition System,简称FRS系统)中基于Haar特征和肤色特征的一种面部检测方法:首先用Haar特征对可能存在人脸的图像进行快速的检测,然后用肤色特征消除相应的彩色图像中误判的人脸检测.实验结果表明该系统具有较快的识别速度、较高的人脸识别率和较低的人脸误检率.

The author proposed an efficient Face Recognition System （FRS） based on Haar feature and skin color algorithm. At the first step, a Haar Feature based face detector is applied for grey-scale image to rapidly detect possible face positions in the image. Later, false face detections are eliminated by skin color filtering on corresponding color image. The results show higher face recognition speed, more accurate face recognition rate, and less false face recognition rate.