Centromere Detection of Human Metaphase Chromosome Images using a Candidate Based Method

Accurate detection of the human metaphase chromosome centromere is an critical element of cytogenetic diagnostic techniques, including chromosome enumeration, karyotyping and radiation biodosimetry. Existing image processing methods can perform poorly in the presence of irregular boundaries, shape variations and premature sister chromatid separation, which can adversely affect centromere localization. We present a centromere detection algorithm that uses a novel profile thickness measurement technique on irregular chromosome structures defined by contour partitioning. Our algorithm generates a set of centromere candidates which are then evaluated based on a set of features derived from images of chromosomes. Our method also partitions the chromosome contour to isolate its telomere regions and then detects and corrects for sister chromatid separation. When tested with a chromosome database consisting of 1400 chromosomes collected from 40 metaphase cell images, the candidate based centromere detection algorithm was able to accurately localize 1220 A. Subasinghe and Department of Electrical and Electronics Engineering, University of Sri Jayewardenepura, Nugegoda, Sri Lanka. J. Samarabandu Department of Electrical and Computer Engineering, Western University, ON, Canada. R. Wilkins Health Canada, 775 Brookfield Road, PL 6303B, Ottawa, ON and Canadian Nuclear Laboratories, 20 Forest Avenue, Deep River, ON K0J 1P0. F. Flegal Canadian Nuclear Laboratories, Plant Rd, STN 51, Chalk River, ON, K0J 1J0. Y. Li, J. Knoll and P. Rogan Departments of Pathology and Biochemistry, Schulich School of Medicine & Dentistry, Western University, ON, Canada. centromere locations yielding a detection accuracy of 87%. We also introduce a Candidate Based Centromere Confidence (CBCC) metric which indicates an approximate confidence value of a given centromere detection and can be readily extended into other candidate related detection problems.