Deep Learning Inpainting Model on Digital and Medical Images-A Review

Abstract

Image inpainting is a method to restore the missing pixels on damaged images. Initially, the traditional inpainting method uses the statistics of the surrounding pixels to find the missing pixels. It sometimes fails to read the hidden information to attain plausible imagery. The deep learning inpainting methods are introduced to overcome these challenges. A deep neural network learns the semantic priors and hidden representation pixels in an end-to-end fashion in the digital and medical. This paper discusses the following: 1) The difference between the supervised and the unsupervised deep learning inpainting algorithm used in medical and digital images. 2) Discusses the merits and demerits of each deep learning inpainting model. 3) Discusses the challenges and solution for the deep learning inpainting model. 4) Discusses each model's quantitative and qualitative analysis in the digital and other medical images