Image Reconstruction Using Priority Patch Algorithm

In this paper, a simple Priority Patch Algorithm (PPA) is presented for exemplar image completion. This novel algorithm is developed for identifying a patch of picture from the surrounding pixels of an image and it fills the unknown regions using the technique of priority patch -matching. Quality picture is required by the people, whenever they take a photograph. Unfortunately, if large unwanted object hides the necessary background image, then it can be filled by the known surrounding pixels to enhance the image. The matching score is evaluated by selecting few coefficients from decomposed frequency coefficients. A simple priority patch algorithm is developed to fill the unknown pixels in a patch image block. The Priority Patch Algorithm is based on a search-array technique, and can be implemented very efficiently. The search-array technique runs well on the Graphics Processing Units (GPU) and good experimental results are obtained on real photographs.