Study On Restoration Algorithm Of Defocusing Iris Image

With the rapid development of information technology, the research on information security has become one of important topics. The identity recognition based on the biology characteristics of human body becomes more important. The iris recognition is a kind of the biometrics technologies based on the physiological characteristics of human body, compared with the feature recognition based on the fingerprint, palm-print, face and sound etc, the iris has some advantages such as uniqueness, stability, high recognition rate, and non-infringing etc. Hence the iris recognition technology has recently become popular in identity recognition.Iris collection is more important in iris recognition. The iris recognition system which is proposed by Daugman has already mature greatly. But the precondition is that recognition has to process when the iris image is clear. Because the great mass of the image sampling equipment don't have the function which can adjust the focus, the image that we gained is often defocusing. The blurry image will bring the difficulty in iris collection, so how to get the clear image is become very important. The study of iris image restoration means a lot for the improvements of the speed and accuracy of subsequent recognition algorithm.This paper aim at the study of blurred iris image. We restore the blurred iris image which is acquired by changeless focus camera.First, we analyzed the reason that iris image is out of the focus. Owing to the variable focus equipment is very costly, now most iris image sampling equipments are fixed-focus. The image is often out of focus, for example, the out of focus because of the relative motion between a person and a camera; when depth of field is small, the object of different depth in imaging area will cause out of focus of different degree. It is an unexpected factor for out of focus to cause the faintness of iris image, the defocusing distance are always unpredictable, which brings the difficulties in defocusing iris image restoration. Something else, there is noise during the image sampling, which effects the quantity of image. Because the information of noise is uncertain, the problem of denoising is also a significant problem for iris image restoration. The key element of iris image restoration lies in the setting of degradation model of blurry image, the problem of denoising and the decision of the parameter of degradation model. The parameter of degradation model in defocusing iris image can be abstracted as a disc function,its corresponding radii of the disc is the only the parameter of degradation model ( the blur radii ). The writer gives a detailed illustration about the getting of this parameter.Second , the writer issues the autofocusing technique for iris image based on the power spectral. The focus is that the changing trend of the power spectral along with the blur radii, between which if there is a function, then the relationship between defocusing distance and power spectral, so as to judge whether the system is in focus by the usage of the power spectral.Third, we used the finite ridgelet transform for iris image denoising. As the preexistence of the ridgelet transform, the wavelet transform has the character of both time domain and frequency domain, it is the same with the denote of transient semaphore. It is ascensive compared with the Fourier analysis, so the application of wavelet is very extensive. Wavelet transform can figure the zero-dimensional or punctate singularities commendably. When the edge character is high-dimensional, wavelet can not describe the character of the picture. But the ridgelet transform, which is put forward on the basis of wavelet transform, can disposal the beeline on high-dimensional space or the problem on hyperplanes. Ridgelet transform can not only express the functions widely with the superposition of ridgelet functions, but also have the rim of ridgelet functions which based on discrete transform and approximate to the orthogonality. It can simulate the practical semaphore with the asymmetry, sequentially, ridgelet transform can describe the character of the picture more accurately and help us gain the better result. Ridgelet transform equals to apply the wavelet transform on the slice of the Radon transform. By using the finite Radon transform, for every finite Radon slice which direction is k , we apply the discrete wavelet transform. Choosing every direction, we can find the finite ridgelet transform. This paper use the finite ridgelet transform to denoising the blurry iris image and find a better result.Finally, the blurred iris image was restored by regularization method. We use this method to restore the blurry iris image in frequency domain. It can close to the exact solutions adequately, and also keep the stationarity of the solutions. Easy to say, regularization method keep the balance of solutions adequately and the stationarity of the solutions. By the experiment, we get a good result.