摘要
提出了一种新的图像距离——图像匹配距离(Image Matching Distance,IMMD).IMMD考虑图像中每个像素与对应图像特定区域的关系,在特定区域寻找与该像素匹配的点,从而将图像的灰度值及其坐标位置引入到图像的相似性度量中.这样使得IMMD对人脸姿态、表情、角度变化具有较好的鲁棒性.用基于图像匹配距离的最近邻分类器进行人脸识别.实验结果表明,基于IMMD的方法优于基于传统欧氏距离和图像欧氏距离的同类型方法.
It presents a new image distance-Image Matching Distance（IMMD）.It considers the relationship between every point of the image pixel and the specific area of the corresponding image,and finds the matching point in this special area,to let the image of the gray value and its coordinates into the similarity measure of image.It makes IMMD have good robustness for the changes of face posture,angle,and the expression.The nearest neighbor classifier is used,based on the IMMD for face recognition.The experimental results show that this method is superior to the method based on Euclidean Distance and Image Euclidean Distance.
出处
《广东工业大学学报》
CAS
2010年第3期64-67,共4页
Journal of Guangdong University of Technology
关键词
图像距离
人脸识别
欧氏距离
图像欧氏距离
图像匹配距离
image Distance
face recognition
Euclidean Distance
Image Euclidean Distance（IMED）
Image Matching Distance（IMMD）
