Efficient pairing of chromosomes in metaphase image for Automated Karyotyping

A Karyogram is an image representation of human chromosomes arranged in order of decreasing size and paired in 23 classes. Automatic pairing of the chromosomes is a difficult task because the chromosomes appear blurred with undefined edges and low level of band pattern information. This paper proposes an efficient approach based on determining the propinquity between the chromosomes to automatically pair them. Besides the features used in traditional karyotyping procedures, a novel Nearness Factor, `N-factor´ is defined to pair the chromosomes based on the similarity in the band patterns. The performance of the algorithm was tested and analyzed using 50 images of group `A´ and group `C´ from publically available data set giving an overall pairing accuracy of 100 % for group `A´ and 97 % for group `C´. The experimentation demonstrates that the proposed `N-factor´ achieves higher and robust performance in pairing chromosomes.