Font Size: a A A

Iris Recognition Based On Log Gabor Wavelets

Posted on:2007-03-27Degree:MasterType:Thesis
Country:ChinaCandidate:S C YuFull Text:PDF
GTID:2178360212957606Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
With the development of information technology and the increasing demand of security, the biometrics recognition based on iris as a high reliability and non-offensive bio-information recognition technology becomes more and more important.Iris recognition system consists of image capturing, iris segmentation, iris normalization and enhancement, feature extraction and matching modules. Based on the recent advancements in iris recognition, a recognition approach using the technique of image processing, signal processing and pattern recognition is presented in this thesis. Some new algorithms and experiment results are also described.In iris segmentation stage, the algorithms of Canny edge detection and Hough transform are improved to detect the outer and inner boundaries of iris. The improved Canny edge detection and block Radon transform are used to isolate the eyelids. For eliminating eyelashes, the threshold technique is used. The commonly used method of fixed areas excluding method is also studied to compare with the method proposed.In normalization and enhancement stage, the homogenous rubber sheet model is used to deal with scale inconsistencies. The region inside the outer boundary of iris is enhanced using histogram equalization.In feature extraction and encoding stage, corresponding to the speciality of the distribution of iris texture, more features are extracted in the inner iris region. Considering the shortage of Gabor function, 1D Log Gabor filters are used to filter the iris texture features in the space-frequency domain, and the phase data is also extracted. Then the weighted Hamming distance is employed for classification of iris template. Finally, lots of tests are carried out to find the optimum values of the parameters.In the classification respect, the minimum distance classifier combining with the threshold with minimum error ratio are used as the decision rule. The experiments implemented on CASIA iris database V1.0 show that, the system performs very well, with a success rate of 96.6% in location and the equal error rate of 0.23% in recognition.
Keywords/Search Tags:Iris Recognition, Iris Segmentation, Feature Encoding, Log Gabor Wavelets
PDF Full Text Request
Related items