DocumentCode
3645523
Title
Automated algorithm for processing of FISH images
Author
Marijeta S. Slavković;Milorad P. Paskaš;Dubravka R. Jevtić;Branimir D. Reljin;Nikola B. Reljin
Author_Institution
Inovacioni Centar Elektrotehnickog Fak. u Beogradu, Belgrade, Serbia
fYear
2011
Firstpage
623
Lastpage
626
Abstract
Assessment of HER2, gene located on the centromere of chromosome 17, provides important information about breast samples since it can be a sign of a potential breast cancer. One of the most widely used methods for measuring HER2 gene copies is the evaluation of FISH images. Analysis involves observing the ratio between HER2 (red) and 17 CEP (green) dots within the nuclei of each cell. In this paper we present algorithm for automated evaluation of FISH images, ie. counting of red and green dots in cells whose nuclei do not overlap. The algorithm includes two stages: nuclei detection and dot detection. Detection of nuclei consists of localization and delineation of nuclei in the blue channel. Detection of dots is based on the mathematical morphology followed by dot counting to determine the HER2/CEP 17 ratio.
Keywords
"Marine animals","Fluorescence","Electronic mail","Proteins","Breast cancer","Manuals","Morphology"
Publisher
ieee
Conference_Titel
Telecommunications Forum (TELFOR), 2011 19th
Print_ISBN
978-1-4577-1499-3
Type
conf
DOI
10.1109/TELFOR.2011.6143625
Filename
6143625
Link To Document