Karyotyping of Human Chromosomes Using M_FISH Images

Mumthas T K; Lijiya A; V K Govindan

doi:10.24297/ijmit.v4i2.4621

Authors

Mumthas T K Department of Computer Science and Engineering, National Institute of Technology, Calicut, India
Lijiya A Department of Computer Science and Engineering, National Institute of Technology, Calicut, India
V K Govindan Department of Computer Science and Engineering, National Institute of Technology, Calicut, India

DOI:

https://doi.org/10.24297/ijmit.v4i2.4621

Keywords:

Chromosome analysis, M-FISH, Watershed Segmentation, Bayes Classifier.

Abstract

Karyotyping is a technique used to align the chromosomes in the decreasing order of size so that the structural and numerical changes can be easily identified. Traditional chromosome analysis is simplified by the introduction of M_FISH (multiplex fluorescence in-situ hy-bridization), a combinatorial labeling technique in which each chromosomes appears in distinct color. In this paper, Bayes classification of M-FISH chromosome images using the features intensity in five channel as well as size of the chromosome is presented. Watershed transform is employed for segmentation . Also reclassification of small misclassified segments to the neighbor region is performed as post-processing to improve the performance. The classifier was trained and tested on a set of images from the dataset [2] and an overall accuracy of 91.09% was obtained.