2 nd International Conference “From Scientific Computing to Computational Engineering” 2 nd IC-SCCE Athens, 5-8 July, 2006 © IC-SCCE MICROARRAY IMAGE ENHANCEMENT TECHNIQUES USING THE DISCRETE WAVELET TRANSFORM Emmanouil Athanasiadis 1 , Dimitris Glotsos 1 , Antonis Daskalakis 1 , Panagiotis Bougioukos 1 , Spyros Kostopoulos 1 , Pantelis Theocharakis 1 , Panagiota Spyridonos 1 , Kalatzis Ioannis 2 , George Nikiforidis 1 and Dionisis Cavouras 2 1 Medical Image Processing and Analysis Group, Laboratory of Medical Physics, University of Patras, 26500 Patras, Greece. e-mail: mathan@upatras.gr , web page: http://mipa.med.upatras.gr 2 Medical Image and Signal Processing Laboratory, Department of Medical Instruments Technology, Technological Institute of Athens, Greece. e-mail: cavouras@teiath.gr , web page: http://medisp.bme.teiath.gr Keywords: Microarray, wavelet. Abstract. The objective of this work was to perform a comparative evaluation of five different wavelet-based filtering techniques in the task of microarray image denoising and enhancement. Clinical material comprised microarray images collected from the Oak Ridge National Laboratory. Image processing was performed in two stages: In the first stage an Exponential Histogram Equalization filter was applied in order to increase the contrast between spots and surrounding background. In the second stage, five wavelet-based image filters (Simple Piece-Wise Linear Mapping Filter (SPWLMF), Hard Threshold filter (HTF), Wavelet Enhancement with Noise Suppression filter (WEWNSF), Garrote Wavelet Threshold filter (GWTF) and Sigmoidal Non-linear Enhancement filter (SNLEF)) were implemented for denoising and enhancing gene microarray spots. The enhancing effectiveness of the five filters was assessed by calculating the Mean-Square-Error (MSE) and the Signal-to-MSE ratio. Results showed that the image quality of the processed images was superior to that of the original images. Significant noise suppression was accomplished by the SPWLMP filter, which scored the minimum MSE and the maximum Signal-to-MSE ratio. Processing time was less than 3 seconds for 512x512 sample images. Wavelet-based processing of microarray images was found to enhance microarray images effectively, by improving the visualization of spots and by suppressing image noise. 1 INTRODUCTION Image processing is an area that wavelet-based techniques have proven to perform successfully. Microarray imaging is considered to be an important tool in bioinformatics. The main benefit of this technique is that it can observe thousands of genes simultaneously. The identification of the genes is closely related with the identification of spots. Due to various sources of noise during image processing [1], the outline of each spot is irregular and, thus, the mean intensity measurements are not accurate [2]. Additionally, the location of the arrayer, as well as sub-arrays contained within the main grid, may vary from image to image. This is due to imperfections during the construction of the arrayer. Moreover, contamination could affect measurements during the scanning procedure [3]. Several methods have been introduced in the past. Statistical methods that include analysis of variance have been introduced by Kerr [4], ratio distribution by Chen [5] and Ermolaeva [6], and Gamma distribution by Newton [7]. All these methods deal with measurement errors, such as cross hybridization. However, the effect of noise has not been previously dealt with. In the present study, a systematic evaluation of five wavelet-based noise suppression filters was performed regarding the enhancement of microarray images.