基于支持向量机的彩色图像人脸检测

Face detection in color image based on support vector machines

本文提出了一种彩色图像中的人脸检测方法,该方法先通过综合定位皮肤区域的一维Haar小波变换和幅度投影来进行特征判别分析(DFA),然后再利用支持向量机(SVM)分类器和贝叶斯统计分析法来区分人脸和非人脸。实验结果表明,该算法鲁棒性强,正确检测率高。

This paper presents a face detection method in color image.Our method derives a DFV by 1-D Haar wavelet transformation and amplitude projections for combining the located skin region.Then the SVM classifier together with Bayesian statistical analysis procedure is classified face and nonface.Experimental results show that the proposed algorithm has strong robustness and high detection rate.