Blurred Image Restoration with Unknown Point Spread Function

Abstract

Blurring image caused by a number of factors such as de focus, motion, and limited sensor resolution. Most of existing blind deconvolution research concentrates at recovering a single blurring kernel for the entire image. We proposed adaptive blind- non reference image quality assessment method for estimation the blur function (i.e. point spread function PSF) from the image acquired under low-lighting conditions and defocus images using Bayesian Blind Deconvolution. It is based on predicting a sharp version of a blurry inter image and uses the two images to solve a PSF. The estimation down by trial and error experimentation, until an acceptable restored image quality is obtained. Assessments the qualities of images have done through the applications of a set of quality metrics. Our method is fast and produces accurate results.