Shyni Carmel Mary S.
Research Scholar, Department of Computer Science, Institute of Distance Education, University of Madras, Chennai, India
S. Sasikala
Assistant Professor, Department of Computer Science, Institute of Distance Education, University of Madras, Chennai, India
Automatic detection and classification of abnormal medical images are very challenging in computer assisted identification of anomaly which helps the physician and the experts. The work presented in this paper attempted integrated method for automatic classification of spinal cord tumor by determining feature values of the Sample image. The integration of algorithm such as Gray Level Co-occurrence Matrix (GLCM) with Multivariate Support Vector Machine (MSVM) and K-Nearest Neighbour (KNN) classifiers approaches are producing effective results in spinal cord tumor classification. In the feature extraction stage, Gray Level Co-occurrence Matrix (GLCM) is used to compute the discriminative features. In the classification stage, the obtained features provide as input for the classification algorithm. Both approaches will classify the abnormal images along with its three types which are based on the location of the tumor existence in the spinal cord in an automatic process. Features extracted with GLCM integrated with MSVM produced 96% accuracy results. Similarly GLCM combined with KNN produced 86.5% accuracy during the classification. The performance shows the efficiency and adeptness of the integrated model.
Keywords
Spinal Cord Tumor, Classification, Gray Level Co-occurrence Matrix, Multivariate Support Vector Machine, K-Nearest Neighbour
Full Article available at https://dx.doi.org/10.20319/mijst.2019.52.2034