基于Gabor小波的人脸识别系统的实现

给出了一种基于Gabor小波的人脸识别系统的实现.该方法先利用模板匹配法定位人脸图像的眼睛,然后经过标准化使图像具有相同的基准,其余特征点的搜索采用弹性图匹配方法;特征点的粗略位置根据训练集的人脸结构知识定位,然后使用弹性图簇的数据结构求得特征点的精确位置,当各特征点的精确位置定位后,人脸的Gabor Jets就可求得;样本图像的相似程度可用与人脸特征点对应的GaborJets作度量函数表达.

An Implementation of a Face Recognition System Based on Gabor Wavelets

In this paper, an implementation of a face recognition system based on Gabor wavelets is presented. First, the eyes of a face image are located by using template matching, then a normalization process is used to produce similar images. The other features are searched in Elastic Bunch Graph Matching. The coarse positions of the feature are located by the training face structure knowledge, and then we use a data structure named Elastic Bunch Graph to adjust the accurate positions. After the accurate positions of the features are located, Gabor jets are extracted from the face image. Similarity between novel images is expressed as function of similarity between localized Gabor jets corresponding to facial landmarks.