摘要
该文提出一个通用的人脸图像识别系统 GFRS[1 ] ,对于人脸库中每个待识目标只存储一幅图像的情况下 ,可以实现各种干扰条件下人脸图像的识别 .该系统提出使用双属性图来表示人脸图像 ,首先利用逐步求精定位法得到人脸图像各局部特征点的位置 ,图中每个特征点由两个特征属性来描述 ,即局部主成分特征系数和 Gabor变换特征系数 ,从而构造了双属性图 .然后给出双属性图匹配函数。
The paper presents a general face image recognition system GFRS [1] . It is expected to recognize the face images subject to vary of disturbances, with the assumption that there exists a single image for each object in data bases. In the system, a dual attribute graph is proposed to represent the face image. First, a location algorithm (refining step by step) is used to detect the location of local facial feature points in the image. Each feature point is described by two attributes of the feature, local principal component coefficient and Gabor transform coefficient. Therefore, the dual attribute graph of the face image is constructed. Then, the matching function and matching algorithm of dual attribute graphs, and the corresponding recognition method are given.
出处
《计算机学报》
EI
CSCD
北大核心
2001年第7期764-769,共6页
Chinese Journal of Computers
基金
中国科学院"九五"基础性研究重点项目 (KJ95 2 -J1-70 7)
关键词
人脸图像识别系统
双属性图表示
GABOR变换
计算机
face recognition, attribute graph matching, facial feature location, PCA, gabor transform
