Font Size: a A A

Studies Of Iris Recognition Based On Wavelet Transform

Posted on:2007-09-29Degree:MasterType:Thesis
Country:ChinaCandidate:F ZhuFull Text:PDF
GTID:2178360185975640Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the rapid development of the economy and the coming of information times, the demands for information security are more and more important. So the identity recognition based on the biology characteristics of human body was paid more atten- tion to. Among various biometrics features of human body, iris has some advantages such as uniqueness, stability and no-infringing. So iris recognition technology is considered as one of the most promising biometrics.In this paper, the advantages of iris used as recognition of identification are discussed and the key technology of the recognition system is summarized. Then several mature iris recognition systems are analyzed. Based on the previous work of other researchers, the following researches combined with wavelet theory are put forward.Firstly, at the iris image pre-processing stage, an improved algorithm of iris location based on two-step method was proposed: First, the inner boundary is found using binary image and the outer boundary is found using histogram equalization image. Then, the improved Daugman operator is used to carry on the accurate localization of the iris image. The unitary iris image is carried out through the transformation from Cartesian coordinates to polar coordinates.Secondly, an improved algorithm of iris recognition based on the wavelet transformation (WT) zero-crossing representation is proposed: appropriate WT is adopted to separate unitary iris texture. Feature extraction is realized by feature quantification, then the Hamming distance is used to match two iris codes.Finally, an algorithm of iris recognition based on the Symlets wavelet is put forward: features are extracted through Symlets wavelet and multiresolution analysis of iris texture. Then the Euclid distance is used to match two iris codes.Experiment result shows that the location method has satisfactory performance both in speed and in precision. In the iris recognition algorithm based on WT zero-crossing, the size of an iris code is reduced and recognition accuracy is better. As for iris recognition based on Symlets wavelet, it is better than that of using Daubechies 4 (db4).
Keywords/Search Tags:iris recognition, iris location, Symlets wavelet, feature extraction, WT zero-crossing
PDF Full Text Request
Related items