分块双向二维主成分分析与模糊分类的掌纹识别

摘要

掌纹识别是一门新兴的生物特征识别技术。提出基于分块双向二维主成分分析（M（2D）2 PCA）和模糊分类的掌纹识别方法。该算法利用M（2D）2 PCA提取掌纹的局部特征，并利用模糊分类策略。这种方法可以有效提取掌纹的局部特征，并且直接对子图像矩阵进行特征抽取，能够精确计算协方差矩阵的特征向量；分类阶段引入模糊理论，应用于掌纹识别问题。最后使用北京交通大学掌纹数据库进行识别实验，结果表明，该方法可得到更高的识别率和更少的识别时间。%Palmprint recognition is an emerging biological feature recognition technology.We present a palmprint recognition method, which is based on blocking bi-directional two-dimensional principal component analysis (M(2D)2PCA)and fuzzy classification.The algorithm uses M(2D)2 PCA to extract palmprint local features,and uses fuzzy classification strategy as well.This method can effectively extract palmprint local features and directly extracts the feature of sub-image matrix,it can accurately calculate the eigenvectors of covariance matrix;in classification stage it introduces fuzzy theory and applies it to palmprint recognition problem.Finally,we use the palmprint database at Beijing Jiaotong University in recognition experiment.Results show that this method achieves higher recognition rate and less recognition time.