掌纹识别中的特征提取算法综述  

Survey of Feature Extraction Algorithms in Palmprint Recognition

作　　者：吴介[1] 裘正定[1] 

机构地区：[1]北京交通大学信息所,北京100044

出　　处：《北京电子科技学院学报》2005年第2期86-92,62,共8页Journal of Beijing Electronic Science And Technology Institute

摘　　要：特征的选择左右着识别的结果,因此特征提取在模式识别中一直占据着重要的地位,如何有效地在高维特征空间中选取一定特征进行分类和匹配是掌纹识别的关键。本文结合近年来发表的文献,按照描述特征的方式进行分类,总结了掌纹识别中的特征提取方法,并对每种方法中的主要算法进行了分析,最后对各种方法的性能进行了比较和展望。Feature extraction plays an important role in pattern recognition. How to effectively extract features from high resolution images to complete the classification and matching is widely concerned. Reviewing the papers published in recent years, the palmprint features according to the method they are described are classified, the feature extraction methods which are used in palmprint authentication are overviewed and concluded and the algorithms used are analyzed and compared. The palmprint feature extraction field is predicted.

关 键 词：掌纹特征提取 结构特征 统计特征 空域-频域变换 子空间 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]