GRENZE International Journal of Computer Theory and Engineering
Authors:
Sreelekshmi A.N
Volume:
2
Issue:
1
Grenze ID:
01.GIJCTE.2.1.524
Pages:
25-29
Abstract
This work aims at the study and implementation of occlusion removal. Occlusion
is a block which disturbs eyesight. In order to get a better clarity we remove the block. It
involves removing the block and filling the gap left behind within the image. There are so
many varieties of algorithms. First, segment the image into structure and texture image.
The sparse technique is used for decomposition of the image and reconstruction of the
structure image. Texture image is reconstructed based on anisotropic diffusion. Finally, the
structure and texture image is recombined and we get an image where occlusion is filled.
This paper includes texture synthesis, exemplar based method and hybrid method for
occlusion removal, sparse method, PDE method. The results are found better since the
patches selected are from the source region, and also regularization is carried out during
diffusion.
This paper aims at the study of various types of inpainting techniques, and how efficiently
they work.