Microarray technology is a relatively new tool in the genetics which allows the researchers to study the behavior of thousands of genes in a concurrent manner. Each microarray experiment results in a number of large images. Compression of microarray images is of importance because many of them are produced, each of them requires large storage space, and they are shared via special data bases. In this thesis a number methods are proposed for the lossless compression of microarray images. For segmentation of images into two regions of spots and background, an adaptive thresholding method is proposed. Also presented are two schemes based on context based modeling. In the first scheme, in addition to classifying pixels into two separate regions of spots and background, the edges of the spots are segmented as a third region. The second scheme is based on adaptive finding of proper contexts in the image. The simulation results have proven the superiority of the proposed method over the standard lossless algorithms as well as superiority over dedicated microarray compression routines.
