A Study On The Noise Detection In Iris Recognition

With the development of information technology in contemporary society,the demands for security are becoming more and more important.So the technology of biometrics identification was improved and applied on many fields.As iris identificationis one of potential biometrics technologies,it has been gradually applied to safety fields.Iris identification is based on iris texture to identify persons,and it is one of most accurate biometrics technologies.Because the iris recognition is in the ascendant abroad,and indomestic it is just on the beginning.So,it is important and valuable to analyze processingthe iris imaged soundly.By boserving the extensive xeperments,the noise of the iris region are mainly composed of eyelid,eyelash and reflection.The different to detect these noise in this paper.1)The eyelid noise contains the upper eyelid and lower eyelids,a combined-line approach was presented for accurate eyelid localization based on Hough Transforms.the traditional method that combins horizontal edge detection with Sobel operator and iteration thresholding,Binary images center for the pupil to be divided into four parts and parabolic Hough transform are used to find the upper and lower eyelid.2)the eyelashes near the lower eyelid is sparsity,short,and turn out,very few projectors in the region of the iris,but the eyelashes near the upper eyelid in the iris usually projectors part of the larger region.Based on these characteristics,eyelashes are detected by the intensity threshold method and real part of log-gabor wavelet.intensity threshold threshold algorithms for detection of eyelashes was presented by calculating the average of grey scale in two rectangle areas,which were near the left and right of the pupi.log gabor wavelet algorithms detects eyelashes by using the real part of log-gabor wavelet,filtering in three directions of each iris image and then adding up filtering images.3)Reflection is divided into strong reflection and week reflection according to the gray value.In order to avoid light effects,threshold selection based on the average gray-scale in each iris image.Experimental results show that the approach is helpful to detect the three noise.