摘要
图像中人脸的检测有广泛的应用范围，提出了一种基于差值图像及子波变换的人脸检测方法．首先根据运动信息从图像序列中检测出候选人脸，然后使用子波变换方法分析候选人脸的直方图，确定出４ 个阈值，再根据此４ 个阈值对候选人脸进行多重阈值二值化，并根据有关人脸的先验知识，在多个二值化图像中定位眼、嘴、鼻．实验结果表明，提出的人脸检测方法能较好地检测出图像中的人脸．此方法在人机交互。
A face detection and recognition system is useful in many applications.In this paper,a face detection method based on image difference and wavelet transform was presented.Firstly,a candidate face was detected from the motion image sequence;then four thresholds were derived from the face histogram by wavelet analysis,the candidate face was transformed to binary images by the four thresholds;finally,the eyes,mouth,nose were located from the binary image.The experimental results demonstrate the efficiency of the algorithm.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
1999年第11期1381-1385,共5页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金
上海市教委科技发展基金
