An Efficient Approach for Medical Image Classification using Association Rules

Medical images are crucial in diagnosis, therapy, surgery, reference, and training. The availability of Digital radiology equipment availability has ensured that digital medical images management gets more attention now. This paper presents an automatic classification system for Computed Tomography (CT) medical images are presented in this paper. In the presented methodology, Bi-orthogonal spline wavelet was used to extract features from the brain, chest and colon CT scan images. Association Rule Mining (ARM) was used for feature reduction leading to the selection of attributes with respect to class label based on frequent sets. On feature selection the CT images are classified using Naive Bayes and k-Nearest Neighbor algorithm. The classification accuracy obtained was compared with and without feature selection using Association Rule Mining.