Date of Award
2010
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Electrical and Computer Engineering
Supervisor
Dr. Jagath Samarabandu
Second Advisor
Dr. Peter Rogan
Third Advisor
Dr. Joan Knoll
Abstract
With the increasing use of Fluorescence In Situ Hybridization (FISH) probes as markers for certain genetic sequences, the requirement of a proper image processing framework is becoming a necessity to accurately detect these probe signal locations in relation to the centerline of the chromosome. Automated detection and length measurements based on the centerline relative to the centromere and the telomere coordinates would highly assist in clinical diagnosis of genetic disorders and thus improve its efficiency significantly. Although many image processing techniques have been developed for chromosomal analysis such as ’’karyotype analysis” to assist in laboratory diagnosis, they fail to provide reliable results in segmenting and extracting the centerline of chromosomes due to the high variability in shape of chromosomes on microscope slides. In this thesis we propose a hybrid algorithm that utilizes Gradient Vector Flow active contours, Discrete Curve Evolution based skeleton pruning and morphological thinning to provide a robust and accurate centerline of the chromosome, which is then used for the measurement of the FISH probe signals. Then this centerline information is used to detect the centromere location of the chromosome and the probe signal location distances were measured with respective to these landmarks. The ability to accurately detect FISH probe locations with respective to its centerline and other landmarks can provide the cytogeneticists with detailed information that could lead to a faster diagnosis.
Recommended Citation
Subasinghe Arachchige, Akila Mike, "Image Processing Techniques for Detecting Chromosome Abnormalities" (2010). Digitized Theses. 4500.
https://ir.lib.uwo.ca/digitizedtheses/4500