PROJECT TITLE :

Blind Deblurring of Natural Stochastic Textures Using an Anisotropic Fractal Model and Phase Retrieval Algorithm

ABSTRACT:

It has been thoroughly researched for natural photographs the tough inverse problem of blind deblurring Edge-type structures or similarity to smaller patches in the image are used by existing techniques to estimate the correct blurring kernel. Natural stochastic textures (NSTs), on the other hand, do not perform well with these approaches because they lack clear edges and contours. Even the tiniest of kernels can have a devastating effect on visuals in NST. Restoration, on the other hand, is a monumental undertaking. An anisotropic fractal model is used to estimate the power spectral density of the blur kernel in this paper. Phase retrieval algorithm adaptations for sparse signals are used to estimate the final kernel. To get even better results, we apply additional limits that are particular to blur filters. Blind deblurring methods reported recently are compared to these results.