基于改进Adaboost算法的人脸检测

Face Detection Based on Improved Adaboost Algorithm

针对传统Adaboost算法存在样本训练耗时、误检率较高的问题,提出了一种基于改进Adaboost算法的人脸检测方法。首先,利用肤色分割对待检测人脸图像进行预处理,然后,对传统Adaboost算法进行改进,在训练分类器过程中加入相关性判断,以减少弱分类器数量和提高样本训练速度,最后,对预处理过的待检测区域用改进的Adaboost算法进行人脸检测。实例测试表明,该方法在人脸检测中能够获得较高的检测率和较快的检测速度。

To solve the problems of time - consuming in training samples and high false detection rate in face detection in traditional Adaboost algorithm, a method of face detection based on improved Adaboost algorithm is proposed. Firstly, the image is preprocessed by skin color segmentation. Then, the traditional Adaboost algorithm is improved by putting relevance judgment in the process of training classifiers to reduce the number of weak classifier and to improve sample training speed. Finally, the image preprocessing is tested by improved Adaboost algorithm again. The experimental results show that the face detection method produces higher detection rate and speed.