Denoising techniques for low intensity fluorescense signal of DNA microarrays

DNA microarrays play a very important role in clinical diagnosis to find out the expression levels of large number of genes simultaneously and mutation of the genes in a given organism. This raw microarray fluorescence signal has a lot of noise, which may originate from auto fluorescence, background noise, non specific binding and detector itself. In this paper, we tried to exploit the different denoising algorithms based on the diffusion filtering and perform their comparative study to find out the best method suitable for low intensity images of microarray. A quantitative measure of image is provided by signal to noise ratio (SNR) of image and visual quality of denoised images compared with universal quality index. This helps to reduce the quantification errors in calculating the expression levels in individual microarray spots and for better representation of the microarray data for clinical diagnosis.