Open AccessMedical Image Inpainting Using Multi-Scale Patches and Neural Networks Concepts
Author(s) -
Nikolay Gapon,
Viacheslav Voronin,
Roman Sizyakin,
D Bakaev,
Arina A. Skorikova
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/680/1/012040
Subject(s) - inpainting , artificial intelligence , computer vision , computer science , image (mathematics) , homogeneous , artificial neural network , scale (ratio) , image restoration , pattern recognition (psychology) , specular reflection , image processing , mathematics , geography , cartography , physics , combinatorics , quantum mechanics
In this paper, we consider the problem of medical images inpainting, where the goal is to reconstruct missing or damaged parts of the image. This is a good tool for medical applications such as vascular restoration, removal of specular reflections for endoscopic images, removal of MRI artifacts, etc. The new method combines the search for patches of various sizes and the operation of a pre-trained neural network. The large patches are used to reconstruct homogeneous areas, and then small patches are used to reconstruct structure image details. As a result, the proposed method provides a plausible of medical images inpainting. Experimental results demonstrate the effectiveness of the proposed method in the tasks of medical images inpainting.