摘要
提出一种基于分块离散余弦变换(DCT)和线性鉴别分析的人脸特征提取方法。该算法对人脸图像进行DCT变换,根据图像块位置和能量分布选择不同的DCT高低频分量构建特征向量,再线性鉴别变换降低特征维数,提高特征的鉴别能力,并利用分类器进行特征的分类与识别。人脸库上的仿真结果验证了该方法的有效性。
A human face feature extraction method based on divided discrete cosine transform（DDCT）and linear discriminant analysis（LDA） is proposed.With this method,the human faces divided into different sub-images and transformed with DDCT,and low frequency coefficients are chosen to construct DCT feature vector.LDA is used to extract the ultimate discriminative features.And the features are classified with the feature classifer.Experimental results on face database validate the efficiency of the method.
出处
《科学技术与工程》
2011年第1期59-61,共3页
Science Technology and Engineering
关键词
人脸
分块离散余弦变换
线性鉴别变换
特征提取
human face divided discrete cosine transform（DDCT） linear discriminant analysis（LDA） feature extraction
