DNA microarray images contain spots that represent the gene expression of normal and cancer samples.
As there are numerous spots on DNA microarray images, image processing can help in enhancing an
image and assisting analysis. The mathematical morphology is proposed to enhance the microarray image
and analyse noise removal on the image. This follows an experiment in which the erosion, dilation,
opening, closing, white top-hat (WTH) and black top-hat (BTH) operations were applied on a DNA
microarray image and its results analysed. Noise was completely removed by the erosion operation and
the images were enhanced.
Normalisation is a process of removing systematic variation that affects measured gene expression levels
in microarray experiment. The purpose is to get a more accurate DNA microarray result by deleting
the systematic errors that may have occurred when making the DNA microarray slid. In this paper,
four normalisation methods of Global, Lowess, Quantile and Print-tip are discussed, tested and their
final results compared in the form of Matrixes and graphs. Ideal and real microarray slides have been
used for this project. It was found that the Print-tip normalisation method showed the closest results to
the real result for an ideal microarray slide and it has a straight median line final graph. The Print-tip
normalisation method uses more than one normalization factor that is divided among intervals which
are dependent on the values of the addition of red and green logarithm.