基于Adaboost算法的人脸检测实现

人脸检测是指把人脸从一幅静止的图像或者动态视频中检测出来,并且指出人脸在图像或视频中的大小和位置。目前存在着大量的人脸检测算法,其中Adaboost算法是比较实用的人脸检测算法。Adaboost算法中人脸的特征采用的是矩形特征,在大量的样本集中,提取样本的矩形特征进行训练,生成多个弱分类器,然后合并多个弱分类器形成一个强分类器,最后级联强分类器形成最终的分类器,以此作为人脸检测的依据。文章先介绍了人脸检测技术,采用Adaboost算法生成人脸分类器,最后利用OpenCV来实现人脸检测。

Face Detection Based on AdaBoost

The definition of face detection is that determine whether or not there are any faces in the image and,if present,return the image location and extent of each face.With further research of face detection,the algorithm is also gradually emerge endlessly.In the circumstances of quick speed and high accuracy of face detection required,Adaboost algorithm is appeared.Some important features are selected,from a large number of which are calculated from integral figures.These features form a series of weak classifier,and then some form a strong classifier.Finally several strong classifiers cascade Cascade classifiers.These are the core ideas to the face detection based on Adaboost algorithm.Face detection technology and Adaboost technology are studied and set forth in details in this paper and Adaboost algorithm is used in face detection.In this paper a small face detection system basede on OpenCV is designed.