复杂背景的人脸检测方法

由于原始的Adaboost方法在复杂图片上检测人脸效果不够理想，所以提出了一种能够处理复杂背景图片的人脸检测方法，即基于肤色的Adaboost检测方法。该方法具有肤色分割的检测率高、适应性强和AdaBoost算法检测速度快等优点。首先，通过人脸肤色的统计特征对图像进行肤色分割，得到候选人脸区域；然后使用经过训练的AdaBoost算法级联分类器对候选人脸区域进行检测，最终得到精确定位的人脸。经过实丐令证明，基于肤色分割的Adaboost人脸检测方法比原始的Adaboost方法在鲁棒性上有了很大提高。

Face Detection Method From Complex Background

As the effect of original Adaboost method in detecting human faces in complex pictures is not ideal,a face detection method for image with complex background is proposed,which is an Adaboost method based on skin color detection.This method has advantages such as high detection rate of skin-color segmentation,wide adaptability,fast detecting speed of AdaBoost algorithm,and so on.First,according to the statistical characteristics of facial skin color,the color image segmentation is processed to get a candidate face region;and then the trained Adaboost algorithm cascade classifier is used to detect the candidate face region to get the human face with precise positioning ultimately.The experimental results show that the Adaboost face detection method based on color segmentation improves the robustness greatly comparing with the original Adaboost method.