Restore digital images using deep learning

Abstract

Photo recovery is the process of rebuilding lost or degraded parts of photos and videos. In the case of a valuable painting, this task will be performed by a skilled artist restoring the painting. In the world of information technology, image recovery refers to the application of complex algorithms to replace lost or damaged parts of image data. Image recovery involves removing noise and sometimes algorithms using ideas about denoising, but image restoration is fundamentally a different matter from denoising. Noise areas usually have some information about the original image, but in image recovery, some areas are completely lost in the original image data. The focus of this thesis is to understand the problems related to recovering noisy grayscale images, to study some algorithms to recover noisy gray images, and to focus on understanding deep learning models to recover images at the beginning. The algorithm will be tested with a program that uses the Python language.