Font Size: a A A

The Study And Implementation Of Palmprint Recognition Methods

Posted on:2003-10-26Degree:MasterType:Thesis
Country:ChinaCandidate:N XuFull Text:PDF
GTID:2178360185485710Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
As one of the most important biometrics techniques, the development of palmprint recognition has a significant influence on real world. In this paper, two kinds of feature extraction methods are proposed based on former work of palmprint recognition and the characteristics of inner palm; the design and implementation of these two algorithms as well as the test of performance are given.The current popular technique in palmprint recognition---print feature based palmprint recognition technique is introduced via a review of palmprint recognition research history. After the discussion of the line based and texture energy based palmprint recognition in detail, we analyze the advantages and disadvantages of these two algorithms and suggest an efficient combination of these two ways---two-level palmprint recognition.In this paper, two novel palmprint recognition methods are proposed in order to overcome the drawback of current algorithms. The whole design and implementation of palmprint recognition system including image acquisition, preprocessing, the basic idea and implementation of palmprint feature extraction and automatic palmprint classification and test are given.As the crucial part of palmprint recognition, this paper mainly focuses on the ideas and usefulness and implementation of these two feature extraction algorithms. Eigenpalms based feature extraction is designed according to the rule of in-class distance and between-class distance. Fisher rule is used then to find the best line that suitable for classification, according to which the sample data is projected from a high-dimensional space to a low-dimensional space. That is to say, the best way for classification is found through maximizing the scatter of different classes and minimizing the scatter of the same class.In order to deal with the unknown transformations of samples as a result of preprocessing and improve the system recognition performance further, a palmprint feature extraction method based on template learning in wavelet domain is designed and proposed. This algorithm learns the ideal templates of different classes from palmprint samples, whose parameters are regarded as...
Keywords/Search Tags:biometrics, palmprint recognition, feature extraction, wavelet transform
PDF Full Text Request
Related items