Image Restoration using 3-Dimensional Discrete Cosine Transform

Image restoration is one of the major issues in the field of image processing. Generally, images are corrupted due to the ageing effect. The previous techniques are based on only single domain i.e.., either local domain or non-local domain. This paper present a new technique called “Image restoration using 3-dimensional discrete cosine transform” which comprises of two domains. Initially, smoothing of image is performed with the help of horizontal and vertical difference operators in the local domain and then synthesizes the textures with the help of discrete cosine transform in non-local domain. A split bregman iterative algorithm is developed to make these two domains more tractable and robust. In this paper, the problem of removing text in an image and filling the corrupted regions is dealt with the help of proposed technique. The proposed method achieves significant performance improvements over the existing state-of-art schemes.