Comparative evaluation of a gaussian mixture models and a seeded region growing techniques for the segmentation of microarray images
Σπυρίδωνος, Παναγιώτα Π.
Αθανασιάδης, Εμμανουήλ Ι.
Νικηφορίδης, Γεώργιος Σ.
Τ.Ε.Ι. Αθήνας. Σχολή Τεχνολογικών Εφαρμογών. Τμήμα Μηχανικών Βιοϊατρικής Τεχνολογίας Τ.Ε.
Κάβουρας, Διονύσης Α.
The purpose of the present study was to investigate
and compare the segmentation ability of the
Gaussian Mixture Models (GMM) against the
Seeded Region Growing (SRG) methods in
microarray spots segmentation. A simulated
microarray image, each containing 200 spots, was
produced. An automatic gridding process was
developed in MATLAB and it was applied on the
images for identifying the centers of spots and their
surrounding borders (cells). The GMM, developed in
MATLAB and the SRG algorithms, using MAGIC
Tool software, were applied to each spot separately
for discriminating foreground from background.
The segmentation abilities of the GMM and SRG
algorithms were evaluated by calculating the
segmentation matching factor for each spot. Optimal
segmentation results were obtained by the GMM,
especially in cases where the spot’s mean intensity
value was close to the background. The GMM
technique was found to be an accurate algorithm in
delineating the boundary of microarray spots and,
thus, in discriminating the spot from its surrounding
Technological Educational Institute of Athens