Research On The Algorithms Of Iris Recognition Based On Gabor Transform

With the awareness of security issues enhanced, traditional identification technology can not meet current requirement. Biometrics technology can fit for the need of modern security system. Iris recognition become the focus by its unique feature in Biometrics individual identification area. It will be performed widely in the application future and will wide market development future.At first, author discuss and summarizes the advantages of iris as biometrics characteristic. And then introduced the structure of iris recognition system, analyzed the localization method which is based on gray surface matching and Daugman's iris normalization algorithm. Due to the source picture existing revolved with the translation ,we adopt translation Pretreatment to eliminate the effect by this noise. Extracting and matching technology based on Gabor translate are mainly studied in this paper.1D-Gabor translation belongs to STFT, Compared with wavelet decomposition, which its define window can not expansion and contraction with the frequency. Daugman proposed the 2D-Gabor filter principle .It made 2D-Gabor translation had the feature of wavelet decomposition that large size window has low frequency, small size window has high frequency. The algorithm for Iris recognition based on Gabor transform can perform identification accurately and quickly .This method is one of the classic algorithms for iris recognition. However, choosing Gabor parameters in design is real a problem for Gabor transform. At present, although there are many papers researched on the application of Gabor filter in iris feature extraction, they have no elaborated on the design in parameters of Gabor filter. This paper proposed two methods for parameters designing in Gabor filter which used in iris recognition.Applied the Gabor filters designed by these methods to iris feature extraction in CASIA and our iris database, experimental results show that it can extract effectively the characteristics of iris texture for iris recognition and has high accuracy.