Anisotropic Image Restoration Based on Image Inpainting with Diffusion Enhancement

Reconstructing the damaged images and improving the quality of an image, results in image restoration. Here anisotropic diffusion based iterative inpainting developed to minimise the noise level in the colour images and enhancing the image boundaries, this approach observed on speckle, Gaussian and shot noise. To reduce noise and topological defects from images, 3D- anisotropic diffusion used to decompose the image into high frequencies and low frequencies and protects the image from losing the information, to enhance the image quality, image inpainiting was used. In this process most of the high frequency decomposed sections got damaged with noise and appears as there is information available at those pixels, therefore the complete restoration process was done on all the high frequency decomposed components so this results in achieving better restored images in mean time. The two effects on images can be reduced by the mixed fusion algorithm i.e., noise reduction by using anisotropic diffusion and distance based neighbourhood image inpainting for restoring the damaged parts. So, this results in reconstructing the damaged image and enhancing the boundaries of the image.