基于高斯模型和支持向量机的人脸检测方法

Algorithm Research of Face detection based on Gaussian Model and SVM

针对目前人脸检测方法速度与精度难以兼有的问题,本文提出了一种结合高斯模型和支持向量机的人脸检测方法。先利用皮肤颜色在YCbCr空间的聚类性,对肤色建立高斯模型以分割出可能的人脸区域,再将这些区域输入到支持向量机检测并标记出检测结果。实验结果证明,本文提出的方法检测效果令人满意。

The current face detecting method has problem that it can not detect both quickly and accurately. This paper proposes an improved method using Gaussian Model and Support Vector Machine（SVM）. On the one hand, we use Gaussian Model to segment the possible facial areas for the skin color being clustered well in the YCbCr space. On the other hand,we use SVM to classify and label the result. Tests show that the method proposed in this paper performs good results.