Font Size: a A A

Application Of Fast Iris Localization Algorithm In Iris Recognition

Posted on:2016-06-17Degree:MasterType:Thesis
Country:ChinaCandidate:Y R WangFull Text:PDF
GTID:2308330470453442Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
With the evolution of science and technology, the rapiddevelopment of the Internet era is coming. People pay moreand more attention to the security of their personalinformation. The identification of identity has also arousedgreat attention Identification of biological characteristics is basedon physiological features and behavior features. Nowadays, the mainidentity recognition methods include physiological characteristicsand acquired habit. Physiological characteristics includes fingerprint,palm print, hand shape, face, iris and retina. The acquired habitincludes language, behavior characteristics of gait. The error rate ofIris recognition is the lowest in all the biological feature recognitiontechnology, which is also the most reliable identificationmethod. Because of uniqueness, stability, the universal, acquisition,acceptability, Iris recognition has become one of the importantresearch contents of biological recognition.The composition and significance of iris recognition system werestudied and analyzed in this paper. It includs iris localization, iris cut,normalized, polar coordinate transformation. A fast that focus oninteresting area iris localization algorithm is improved in this paper,including: (1) In order to avoid the loss of the human eye iris information,taking down the original iris image center half part as interest areas.(2) Extract iris boundary point with sobel operator, filter to a fewsmall border, cylindrical iris localization.(3) Take the average pixel value as threshold, the gray level as thepupil when the value less than the threshold value, then the sobeloperator extract the boundary point in the application, implement theiris location within the circle.The algorithm improves the speed of iris localization, comparedwith the methods take a quarter of the area of interest, the advantageis it will not lost the iris texture information.Iris feature extraction, the iris texture extraction with2-D Gaborwavelet transform that can easily extract the texture information ofdimensions and direction. This article points4directions,28vicetexture image from seven scales after Gabor, filter and extract the irisinformation, reduce the influence of illumination changes and noise tosome extent. By using discrete cosine transform feature reducedimension, not only retains the original important characteristicinformation but also overcomes the drawback of traditional algorithm,which enforce the speed of computing.Iris recognition is achieved by using template matching method.This method is implemented by calculating the Iris feature to betested and the Iris feature in sample database. In this paper, thealgorithm of iris recognition is implemented by Matlab language and Clanguage to realize multiple simulation, finally works out thesimulation map.
Keywords/Search Tags:Iris recognition, Hof transform, Iris location, Gabor wavelet, Feature extraction, Template matching
PDF Full Text Request
Related items