An efficient algorithm for automatic classification and centromere detection in G-band human chromosome image using band distance feature

Abstract

In chromosome analysis, centromere is an essential component. By analyzing centromere, genetic disorder can be identified easily. In this paper, automatic classification and centromere detection in human chromosome image using Band Distance Feature is proposed. Initially the microscopic image of G-band chromosome is preprocessed in order to remove the blobs. Then, the image is segmented using labelling algorithm and the endpoints are calculated. Now, the overlapping chromosomes are removed when the number of end points is greater than two. The non-overlapped chromosomes are straightened using Reversible Projection algorithm. From the straightened chromosome band distance feature is calculated. The extracted features are given to the ANN classifier to identify the class of chromosome and to calculate the centromere. From the experimental results, it is observed that the proposed method is superior to the traditional method.