Iris Recognition Algorithm Based On Support Vector Machine And Hamming Distance

With the progressive development of social economy and the progress of computer network technology, more and more people pay attention to protect the information security and privacy, due to the traditional identity authentication method cannot meet the needs of human has gradually been replaced by the use of biometric identification technology. Iris recognition is one of the most representative biometric technologies, because it has the universality, uniqueness, stability, easy detect ability, high security, good fraud prevention and many other advantages, and is known as the most promising biometric identification technology of the 21 st century.Generally speaking, a complete iris recognition system includes four parts:iris image acquisition, preprocessing image, feature extracting and encoding, feature matching. Based on this, the main work of the paper is around the following aspects:Firstly, image preprocessing: in order to eliminate the influence of high-frequency interference on iris localization; this paper proposes a method that takes smoothing step before iris localization to slove this problom, then makes use of the method which combined with the feature of iris gray and geometric characteristics to locate the iris: according to the principles of geometry, both hypotenuses are diameters of the inside and outside boundaries of the iris, then we can get the parameters of the iris and locate it accurately. Experimental results show that: This method is not only simple to calculate, also high accuracy. Fially,we take adventage of the normalization and enhancement technology to get a texture clear image of iris.Secondly, Feature extraction and coding: using one dimensional Log Gabor wavelet to extract iris texture feature. Specific processing method is: considering the iris image after preprocessing as a two-dimensional signal, then break it down into a series of one dimensional signal, and make the one-dimensional signal respectively with one-dimensional Log Gabor wavelet convolution operation, finally we can acquire the characteristic coding by phase encoding method.Thirdly, Feature matching: according to the traditional iris recognition method focuses on the feature extraction of the phenomenon, the paper proposes a iris recognition algorithm focusing on the identification of a pattern matching, which based on Support Vector Machine(SVM) and Hamming distance of the iris recognition method: after get the iris feature encoding, first feature matching, using the method of SVM for failing to correctly identify or declined to identify the image again through the Hamming distance is used to identify the secondary, and the output of the final recognition results.In order to test the feasibility of the algorithm, this paper provides automation research institute of Chinese academy of sciences of the iris database as samples, and using MATLAB R2010 b simulation as test platform. The final test results shows that: the algorithm of this paper on the recognition accuracy and recognition speed has reached the desired results, and it can provide a theoretical basis for further research in the future.