Font Size: a A A

Research And Implementation Of Iris Recognition Method Based On Morphological Skeleton And Gabor Phase Feature Analysis

Posted on:2019-07-05Degree:MasterType:Thesis
Country:ChinaCandidate:Y C SheFull Text:PDF
GTID:2428330566468731Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Iris Recognition is a highly secure biometric identification technology.The stable and rich texture features of the human eye iris region ideally own the popularity,exclusivity,stability,and accessibility that biometrics technology requires.Compared with the more mature technology such as fingerprint recognition and face recognition,the features of iris recognition are outstanding with good research value and application prospects.The thesis summarizes the advantages and disadvantages of the mainstream iris recognition algorithms.A study and experimental analysis is conducted on the deficiencies in the implementation of mainstream algorithms.Then the solution and designs are given for an iris recognition system with better efficiency and fault tolerance.The research completed by the thesis is as follows:(1)For the two-step positioning algorithm of Wildes,the method of 3D parameter cumulative space voting to locate the inner and outer edges of the iris is too low in calculation efficiency.So an iris edge location and fitting algorithm based on improved Hough transform is proposed.This algorithm uses a non-linear least square method to fit the eyelid margins,and uses the eyelid edge curve as a reference to separate the interference points in the discrete edge pixels.According to the geometric features of the inner and outer edges of the iris,the iris region in the human eye image is segmented using circle fitting and parabola fitting.Experimental results show that compared with Wides algorithm and Daugman algorithm,this algorithm can effectively improve the positioning efficiency and positioning accuracy of standard iris and occlusion iris.(2)For the problem that Daugman feature extraction algorithm has insufficient analysis of the overall features of the iris texture and the local features under different frequencies,a feature extraction algorithm based on morphological skeleton and Gabor filter is proposed.This algorithm Extracting Normalized Skeleton Features of Normalized Iris Using Morphological Algorithm and it uses a multi-frequency,multidirectional 2D-Gabor filter to extract local features.Combined with the iris recognition method proposed in the paper,the intra-and intra-class iris recognition was completed.Experimental results show that compared with Daugman algorithm and Boles algorithm,this algorithm can effectively reduce the false acceptance rate and false reject rate of iris recognition and improve the accuracy of iris recognition.(3)Based on the iris location and iris feature extraction and recognition algorithm proposed above,the C++ language is used as a development language,and the iris recognition system is designed and implemented based on Gdiplus,OpenCV image processing function library and modular development concept.The system includes three main functional modules: iris positioning,iris feature extraction and iris matching.
Keywords/Search Tags:Iris recognition, edge location and fitting, feature extraction, feature matching, Gdiplus
PDF Full Text Request
Related items