摘要
针对人脸检测的特点,深入分析了基于Haar-Like小波特征和AdaBoost算法构造强分类器的人脸检测算法,并将此算法应用于ORL人脸库。在VisualC++6.0平台下,首先利用AdaBoost算法训练得到用于人脸检测的强分类器,然后利用该分类器进行人脸检测。实验表明,提出的算法在保证检测精度的同时,极大的提高了人脸检测速度,且受光照变化影响小。
On the basis of to the character of the face detection, a face detection algorithm based on Haar-Like and Ada- Boost was analyzed. And the algorithm was tested on the standard ORL face database. At the platform of Visual C＋＋ 6.0, firstly, H（x） the for face detection was acquired by using AdaBoost. Then detected face based on the H（x）. The experimental result indicated that the algorithm could get an all-right accuracy with reducing the running time.
出处
《长春理工大学学报（自然科学版）》
2009年第2期248-250,共3页
Journal of Changchun University of Science and Technology(Natural Science Edition)
基金
长春市科技攻关项目(07163UC070)
