摘要
针对由于外貌、表情、肤色等不同,给人脸检测带来很大困难的问题,提出一种基于肤色特征与边缘检测相结合的人脸检测方法,将归一化后的RGB空间转换为YCbCr空间,再对图像进行光线补偿等预处理;在对肤色区域进行判决的过程中,采用自适应阈值的方法。本文算法实验结果证明具有一定的可靠性和有效性,而且计算速度快、方法简单、定位率高、检测效果好。
As the appearance of human face, expression, color and so on are different, these give a great deal of difficulties in face detection, this paper presents a face detection method based on skin-color features and edge detection. Firsly,we convert the normalized RGB space to YCbCr space. Then, we compensate the light of the image. In the decision process of skin-color region,we use adaptive threshold method. The experimeutal results show the validity and reliability of this method, and this algorithm is fast, simple, high orientation and effective detection.
出处
《桂林电子科技大学学报》
2010年第3期212-215,共4页
Journal of Guilin University of Electronic Technology
关键词
人脸检测
肤色特征
边缘检测
自适应阈值
face detection
skin-color features
edge detection
adaptive threshold
