Image Deblurring Algorithm Using ACCELERATED Gradient Descent Method
Sreenivas Sasubilli, Kumar Attangudi Perichiappan Perichappan, Hayk Sargsyan · 2018
In this paper, we address an important problem of current image processing practice in a simply implementable way. More specifically, we develop a parametric image processing algorithm for removing blur effects from an image and recovering the initial unnoised image. Mathematically, the debluring problem is formulated as a non-convex optimization problem which is solved using the accelerated minimization method (accelerated gradient descent). Numerical analysis of several blurred images allows to estimate the advantages and drawbacks of the algorithm.