Automatic segmentation of DNA microarray images using an improved seeded region growing method

Microarray technology has emerged as one of the robust methodology for quantitatively analyzing gene expressions of thousands of genes simultaneously. The experimental design, image processing and data analysis are the three major stages of microarray based analysis. The main goal of array image processing is, to measure the intensity of the spots and quantify the gene expression values based on these intensities. This paper describes segmentation of microarray images using an improved seeded region growing method. The seed and threshold values were selected automatically depending on the characteristics of the spot. Experimental result shows that the method is very effective for segmenting low intensity spots having irregular shape. It is robust against the common noise peaks in the microarray images. The algorithm was implemented using Matlab software. The proposed method has been tested on variety of microarray images obtained from Stanford Microarray Database (SMD).