Numerical analysis of AOR iteration for anisotropic diffusion equation in image blurring

Abstract

Image blurring is one of the most crucial techniques in image processing and computer vision. The equation used for solving image blurring problem is known as anisotropic diffusion equation. It able to smooth the inner region of the image without blurring the significant part such as fine details and sharp edges. The main purpose of this paper is to examine the efficiency of Accelerated Over Relaxation (AOR) iterative method with two weighted parameters for solving image blurring problems. To do the performance analysis based on the number of iterations and computational time, three iterative methods which are Gauss-Seidel, Succesive Over Relaxation (SOR) and AOR had been examined to produce the same quality image of Jacobi iterative method. Based on numerical results of the experiments taken out, the findings have shown that AOR iteration is slightly better than Gauss-Seidel and SOR iteration in reducing the number of iterations and computational time approximately by 30% and 2% respectively.