| Iris recognition technology is a new area of biometric recognition technologies, witch identify people by means of some sole and stable physical features. Iris lies between sclera and pupil like a disk, has complicated texture, witch is fit to identify. Iris has better security and lower error rate than the other biometric recognition technologies. Because the pupils can dilate and shrink, iris can distinguish between the active eye and the imposter. So iris recognition is widely seen as one of most promising biometric recognition technologies.Make use of wavelet transforms to extract the features of iris texture is the emphasis of this paper.The basic theory of wavelet and uncertainty relations are first introduced in this paper, then the features of iris texture and iris image preprocessing are simply described. Iris image preprocessing includes four parts: iris localization, iris normalization, iris image enhancement and choice of the valid region. Solve disturbs from the image excursion, pupil dilation and shrinkage, non-linear illumination and the shelter of the eyelid and eyelash. A new method is presented to choose the valid region of iris image in this paper. This method not only increase the texture of iris, but also easy to resolve the iris rotation.The iris image after reprocessing has clear texture feature. Wavelet transformation with outstanding time frequency partial ability can be used to extract the texture of iris image. 2D-Gabor filters can simulate 2D receptive field profiles of the cortical simple cells, and can reach the lowest limit of uncertainty relations, so 2D-Gabr filters can capture the information of orientation and frequency of iris image. 2D-Gabor filter is taken as the basic function of wavelet, and the parameters of Gabor filter are selected according to the property of system. Because the texture of iris is sparse outside and dense inside, the range of frequency is divided into two bands. To solve the rotation of iris, a new... |
PDF Full Text Request