一种改进的人脸检测新技术

An Improved New Technology of Face Detection

该文提出了一种基于多特征融合的人脸检测新方法。该方法首先用肤色分析算法对图像进行肤色分割．得到人脸区域的初步结果；然后根据灰度特征图像用基于GA的方法找到多个模板的人脸区域：最后用基于Adaboost多模板融合算法形成最终检测结果。通过实验表明，该方法与传统方法相比，识别率得到明显提高。

This paper proposes a new method based on Face detection based on Multi-Features Fusion. it consists threes steps： skin detection using probability-threshold algorithm . localize face region using GA based on multi-model, construct the last result using Adaboost classifier based color and texture. The experiment shows that the new method can provide satisfactory detection performance.