Wavelet Transform In Iris Image Denoising Application

Biological feature identification has been generally applied with the development of communication and information technology, which is based on human face, fingerprint, DNA, iris, etc. especially the iris-based identification draw researchers'more attention. the iris information and some noises by uncertain reasons are gained at the same time. The noises greatly affect the iris-based identification.Wavelet-transform has popularly succeeded in de-noising for its multi-scale, multi-resolution features. In the paper according to the relative features of iris image's wavelet coefficients, wavelet threshold de-noising algorithm, the least square algorithm and the edge detection on the basis of wavelet modulus maximum combine to detect and filter the light spots and the eyelid noises in iris images, the results show the noises in iris image are efficiently eliminated but the edge features are reserved, which lay a good foundation for the successive works, such as the feature extraction, the pattern match of the iris image and the application of the iris recognition system.