Efficient feature extraction and classification of chromosomes

Karryogram is a preliminary procedure to detect the most characteristic signs of a disorder that may require for further investigation of medical applications mainly for cancerous. Diagnosis of karryogram is generally very complex, eroding and a time consuming operation. As of now it requires fussy attention to details and calls for meritoriously and trained personnel. Normally chromosomes are essential genomic information carriers which contains 23 pairs. This paper suggests a efficient classifier Support Vector Machine (SVM) for classifying chromosomes in comparison to the already existing methods such as support vector machine based medial axis and density profiles. The features are extracted based on GLCM (Gray level co-occurrence) feature extraction algorithm which is very well known for its high accuracy. First order features and GLCM features of chromosomes are extracted from the segmented image. As a prerequisite, image segmentation needs to be done by using Fuzzy-C Mean (FCM) procedure to obtain efficient features in coordination with SVM which is used to classify the chromosomes from the available pairs of 23 chromoseomes. Using this methodology increased the accuracy of classification results. Simulation results are carried out in MATLAB to support the analysis.