Abstract: On the basis of weighted 2D Discrete Wavelet Transform(2D-DWT) and Fisher Linear Discriminant(FLD), this paper presents a face recognition algorithm. It decomposes face image into level two by wavelet db2, extracts image features using FLD method for the decomposed image. The extracted features are recognized by nearest neighbor classifier. On the basis of ORL face database, experimental results show the algorithm has high recognition rate.

Key words: 2D Discrete Wavelet Transform(2D-DWT, Fisher Linear Discriminant(FLD), face recognition, feature extraction

摘要： 基于加权二维离散小波变换(2D-DWT)与Fisher线性判别(FLD)，提出一种人脸识别算法。利用db2小波对人脸图像进行2层小波分解，对于分解图像利用FLD法进行特征提取，运用最近邻分类法对提取的特征进行分类识别。在ORL标准人脸图像库上的实验结果证明，该算法能取得较好的识别率。

关键词: 二维离散小波变换, Fisher线性判别, 人脸识别, 特征提取

CLC Number: 

• TP391