摘要
为了能够实现单幅彩色图像中的人脸检测,提出了基于肤色模型和层次滤波技术的人脸检测算法。它通过基于肤色的鲁帮性进行肤色建模,在图像中精确的提取出肤色区域,再经过方向校正后,检验出一系列的人脸特征、像素亮度的变化以及连通区域的空间排列,可以较好地检测出图像中的脸部区域。实验结果表明,该方法可以适用于不同光照条件下的人脸检测,即使在脸部特征没有被很好的检测出的情形下也可以较好完成人脸检测。
In order to accomplish face detection from single color images, a new architecture for face detection in color images based on skin locus and successive detectors is introduced. The detection is based on a robust modeling of skin color, called skin locus, which is used to extract the skin-like regions. After orientation normalization and based on verifying a set of criteria （face symmetry, presence of some facial features, variance of pixel intensities and connected component arrangement）, only facial regions are selected. Results show that our system can detect faces and deal with different conditions （size, orientation, illumination, and complex background）, even when these features are not well defined.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第19期4253-4256,共4页
Computer Engineering and Design
基金
江苏高校自然科学研究计划基金项目(08KJB520001)
关键词
图像处理
肤色模型
层次滤波
人脸检测
最佳匹配椭圆
image processing
skin locus model
hierarchical filtering
face detection
best-fit ellipse
作者简介
赵文东（1972-），男，硕士，副教授，研究方向为计算机图像处理。E-mail：zwd_139@163．com
