Research Of Iris Identification Based On Gabor Filter And Wavelet Transform

With the development of information technology, information security is becoming an important and urgent issue. Iris recognition is a cutting-edge research in the field of pattern recognition, which is a new authentication method based on biometric. It has been widespread concerned by domestic and foreign government departments, academia and industry because of its stability, high reliability and non-contact, and it has been in-depth studied, and it is also considered as one of the dominant biometric technologies in the future.Usually a typical iris recognition system includes iris imaging, living body to detect and iris recognition, and this paper focuses on the last part. On the basis of previous work, this paper discusses on iris image quality evaluation, Gabor filter designing, iris coding, iris recognition matching and so on, and also gives some improvements and innovations. In addition, we know that affected by acquisition equipment, environment and other factors, not all collected iris images are suitable for iris recognition. Based on this, this paper presents an improved iris image quality assessment method. Before the iris preprocessing, identifying all the collected iris images, and thus can discard those low quality iris images which are not suitable for the identification at the beginning. Under the basic premise of not affecting the speed of iris recognition, it can improve the accuracy of iris recognition to some extent.This paper analyzes the iris recognition algorithm based on Gabor filter and wavelet transform. In the iris recognition algorithm based on Gabor filter, combining with iris texture features, this paper designs the Gabor filter parameters which are suitable for iris recognition. Then combining the respective advantages and disadvantages of the two algorithms, this paper proposes a new iris recognition algorithm based on the combination of Gabor filters and wavelet transform. Specifically, for the enhanced iris image, firstly coding feature by the the Gabor filter designed before, then matching by an improved method from coarse to fine proposed in this pape. That is, for the feature coded by Gabor filter, coarse matching using the improved weighted Hamming distance, thus general classifying the iris image and forming a smaller sample set, and then based on the coarse matching, for the feature coded by wavelet transform with the smaller sampleset, fine matching using the improved weighted hamming distance matching method, so achieving the matching strategy from coarse to fine, and improving the accuracy and efficiency of iris recognition. The experiment results show the effectiveness of the algorithm proposed in the paper.