Non local image restoration using iterative method

摘要

This paper proposes image restoration using interpolation, inpainting and denoising the noisy image by iterative method. The main task of image restoration is to capture a noisy image and estimating the original image. The important scheme of image restoration is to stabilize the defects which can corrupt the original input image. In this paper, restoration of image can be implemented by three methods. First, certain level of noise is added to the input image and the noise can be reduced by iterative method. In this, noise level added is made constant and iteration is performed. Second, the damaged image is selected and the quality image is restored by estimating the local properties of the image (inpainting). Third, restoration is done by using interpolation of up and down sampling. For every method, the quality metrics like PSNR and SSIM are calculated. From these PSNR and SSIM values the tabulation and performance graph are obtained. The improved efficiency of image restoration using image denoising, interpolation and inpainting is achieved and they are compared with all other techniques. Convergence time is also reduced when compared with other techniques.