Quality processing of microarray image data through image inpainting and texture synthesis

Image inpainting techniques are usually employed to reconstruct small areas of images that may have been damaged, producing results which are often aesthetically indistinguishable from the original. In this paper these techniques are used in a more error critical manner, as the aim is to reconstruct areas of real-world noisy cDNA microarray images and, thus, help reduce variation and remove background artifacts before post processing of this data. The paper evaluates a number of existing techniques and proposes some directions to further develop them in the context of microarray image processing.