基于小波变换和支持向量机的人脸检测

Face Detection Based on Wavelet Transformation and Support Vector Machines

提出一种新的人脸检测方法———基于小波变换和支持向量机的方法。其方法的新颖之处体现在:通过综合原输入图像的小波变换值、灰度值的投影来进行特征分析;运用统计模型来估计类条件概率密度函数;运用最优的分类方法———贝叶斯分类器进行判决分类。人脸类采用正态分布建模,而非人脸类(包括除人脸类之外的一切事物)仍用正态分布来建模是不合理的。但可以用支持向量机方法从非人脸类中抽出一些跟人脸类很接近的非人脸类的特殊子集,然后对这特殊子集用正态分布进行建模。

Presents a novel face detecting method based on wavelet transformation(WT)and support vector machines(SVM).The novelty of this method comes from the integration of the discrimination feature analysis of the input image,the statistical modeling of face and nonface classes and the Bayes classifier.First,feature analysis derives a discriminating feature vectors by combining the input image's wavelet transformation and its amplitude projections. Second, statistical modeling estimates the conditional probability density functions of the face and nonface classes ,respectively.Finally,applys Bayes classifier to detect.While the face class is modeled as a normal distribution,the nonface class is much more difficult to model due to the fact that it includes 'the rest of the world '.However,one can still derive a subset of the nonfaces that lie closest to the face classes using support vector machines methods,and then model this particular subset as a normal distribution.