Numerical Solutions for Poisson Image Blending Problem using 4-EDGSOR Iteration

Poisson image blending is an operation in image processing to generate a new image by using Poisson partial differential equation. In this paper, 4-EDGSOR iteration is used to solve the Poisson image blending problem and its efficiency in solving the proposed problem is illustrated. The approximatio...

Full description

Bibliographic Details
Main Authors: Jeng, Hong Eng, Azali Saudi, Jumat Sulaiman
Format: Article
Language:English
Published: 2018
Subjects:
Online Access:https://eprints.ums.edu.my/id/eprint/24973/1/Numerical%20Solutions%20for%20Poisson%20Image%20Blending%20Problem%20using%204%20EDGSOR%20Iteration.pdf
Description
Summary:Poisson image blending is an operation in image processing to generate a new image by using Poisson partial differential equation. In this paper, 4-EDGSOR iteration is used to solve the Poisson image blending problem and its efficiency in solving the proposed problem is illustrated. The approximation Poisson equation is formed by applying a finite difference method. Then, the rotated Laplacian operator which is constructed by a rotated finite difference scheme is used in this paper. Then a linear system is formed and solved using the 4-EDGSOR iterative method. The performance of the 4-EDGSOR iterative method in solving the proposed problem is compared to the SOR and 4-EGSOR iterative methods. The results obtained from numerical solutions showed that the 4-EDGSOR iterative method required lesser time and number of iterations to blend an image. From quality point of view, all images obtained the same natural look.