Research And Implementation Of Iris Recognition Algorithms

Biometric recognition technology, which identifies people through their uniqual features, becomes hot recently due to its convenience and efficiency. Human has many biometric features, such as fingerprint, face, eye and iris. Among these, iris has very special structure which makes the identification has high accuracy; therefore, Iris identification technology is regarded as one of the most widely-used biometric identification technology in the future. In this paper, we introduce the below aspects:1. Through analyzing a few existing algorithms, we proposed two new iris image quality evaluation algorithms. Both take account of the image definition and eyelash occlusion problem, and are universal in different iris acquisition system.2. In the iris localization part, to improve the efficiency and accuracy of the proposed algorithm, we proposed a new method based on the iris intensity value, in which we adopt edge points detecting and curve fitting. Then, we proposed another iris localization algorithm based on Multi-resolution Analysis. The most important merit of this algorithm is that it's seldom affected by lash occlusion problem and illumination.3. In the feature extraction and matching part, we analyze the different matcing results when use the Gabor Transform with different parameters. At last, we give a new feature extraction method.4. We introduce the transplanting method of the algorithm to TI TMS320 C6713b DSPs platform and design the embedded system based onμC/OS-II.All the algorithms are proved to be valid accoerding to the experiments on CASIA iris database. Compared with the previous algorithm, the algorithm poposed in this paper improve a lot. Meanwhile, the system is stable and the time cost to process one image is less than 0.2s.