Image enhancement for DNA microarray gridding using Conditional Convolution Sub-Block Histogram Equalization

Abstract

In this paper, a new image enhancement method for DNA microarray images, namely the Conditional Convolution Sub-block Histogram Equalization (CCSHE), is presented. This is accomplished by first, converting the image into a binary image using the histogram transformation function. By doing this, the background and some of the artifacts in the image are separated from the DNA microarray spots. Secondly, this method also applies a conditional convolution filter depending on two input values which control how many times two individual convolution kernels are applied. By performing these operations, the unwanted signals known as artifacts are reduced, providing a more accurate gridding result.