摘要
边缘的提取在人脸识别技术中具有重要的意义。传统的算法检测到的边缘信息往往是不完整的,而且在检测噪声污染图像时会得到许多虚假的边缘;而神经网络算法由于没有样本的压缩功能,导致训练量过大。文章在传统算法的基础上结合神经网络算法对人脸图像的边缘进行检测。实验结果表明,该方法快速稳定,得到的边缘图像边界封闭性好,抗噪能力强。
The abstraction of the edge has important meanings in people＇s face recognition technology.The edge information that the traditional algorithm measures is ofen incomplete and easily gets many false edges when detecting the noise image.Because the neural network algorithm does not have compression function of the sample, so it needs Considering too many trained samples.This paper combines the neural network algorithm that detecting edge of face image of people on the basis of traditional algorithm. The experimental result indicates, this method is steady fast and the edge closed well and with better anti-noise characteristic.
出处
《长治学院学报》
2008年第5期35-37,共3页
Journal of Changzhi University
关键词
人脸识别
边缘检测
神经网络
face recogonition
edge detection
neural network
