Patch-based denoising with K-nearest neighbor and SVD for microarray images

Abstract

Irrespective of certain major advancement in filtering process in medical images, the denoising operation in microarray images are still considered to be unsolved and offers a large scope of research. Existing denoising principles are less investigated on such complex and massive dimensional microarray image that leads to the development of the proposed system. We present a method of performing simple denoising operation considering the presence of Gaussian noise in microarray image. From the target image denoising method, an improved version of patch-based denoising approach has been developed considering various forms of distance-based matching methods. The study outcome of the proposed system has been found to offer better peak signal-to-noise ratio and structural similarity index in contrast to existing filtering techniques.