Share Email Print

Proceedings Paper

Inpainting using neural network in medical image analysis (Conference Presentation)

Paper Abstract

We consider in this paper the problem of image inpainting in medical image analysis, where the objective is to reconstruct missing or deteriorated parts of an image. It is a good tool for such medical applications as vascular reconstruction, specular reflection removal for endoscopic images, MRA artefacts removing etc. Most inpainting approaches require a good image model to infer the unknown pixels. The proposed approach uses the modified exemplar-based technique. A novel approach combines mapping from image patches and pre-trained deep neural network. In our work, we exploit the concept of sparse representation, which takes a group of nonlocal patches with similar textures as the basic unit instead of a patch. Moreover, the color and multi-direction constraints are incorporated into the optimisation criterion to obtain sharp inpainting results. As a result, the proposed method provides plausible restoration while propagating information of edge for the target region. Experimental results demonstrate the effectiveness of the proposed method in the tasks of medical image inpainting.

Paper Details

Date Published: 15 May 2018
Proc. SPIE 10668, Mobile Multimedia/Image Processing, Security, and Applications 2018, 106680K (15 May 2018); doi: 10.1117/12.2305064
Show Author Affiliations
Viacheslav V. Voronin, Don State Technical Univ. (Russian Federation)
Gevorg Karapetyan, Institute for Informatics and Automation Problems (Armenia)
Sos Agaian, The City Univ. of New York (United States)

Published in SPIE Proceedings Vol. 10668:
Mobile Multimedia/Image Processing, Security, and Applications 2018
Sos S. Agaian; Sabah A. Jassim, Editor(s)

© SPIE. Terms of Use
Back to Top