| With the information times coming and the development of the network technology, the traditional safety system which based mainly on token or password can't meet the requirement of modern security system .so people make for biometrics technology which has been the new growth area in the economy. As the important member in this area, iris recognition can be performed widely in the application future and has wide market development future, but the iris recognition is in the ascendant abroad, and in domestic it is just on the beginning. So, it is important and valuable to analyze processing the iris images soundly.In this article the author adopts "Y-cycle" development mode and has done some valuable experiments on the iris images processing and gain good rewards .At first, author discusses and summarizes the advantages of iris as a biometrics characteristic. And then introduce the structure of iris recognition system. Analyze and sum up the characteristics of the iris images in gray space based on the result of observing large number of iris images. It becomes an important basis of the following processing.Iris localization is an important part of iris system. Its speed and precision affect the system performance significantly. A new blurring iris exterior boundary localization algorithm based on the geometric features of circle is proposed because of the difficulties in iris localization. And an algorithm, which determine iris interior boundary according to sum of deviation is proposed too. The experimental results demonstrate that this method is of fast iris location speed and high stability. It also reduces search blindness existed in original method.To lower FAR(false acceptance Rate) and FRR (false rejection rate), the author establish iris image quality evaluation module by considering the main factors which may cause failure in logging on. Three indicators, "image definition ","iris eccentricity" and "iris visibility" were proposed to seek out the "bad" image. For better performance, these three indicators were combined to one-Q. the relation between Q and the rate of the iris recognition was also analyzed.To ensure good quality of the iris image and get rid of the disturbing factors in iris recognition, and analyze different noise in the iris field and take various of the methodsto eliminate the noise, an iris image revise module is built in this article which revise some deviance including in shift, rotate, scale zoom.Gabor function can simulate the figure of receptive fields in mammiferous optokinetic neurons well. 2D gabor function is of good united localization and direction selection in space and frequency fields. So it is very fit for extracting the iris texture. In this article, 2D gabor wavelet transformation is used to treat the iris image and to code iris characteristics.
|
PDF Full Text Request