Medical Image Classification Using Decision Trees

Computer aided Diagnosis (CAD) has been usually applied in MR images to detect tumour and treat diseases. This work is classifying automated system to differentiate the MR images as affected and not affected quickly. It helps to doctor's analysis and decrease the investigation time. This work is therapeutic correction for the pancreatic tumour classification by performing 3 classifiers such as C4.5, C5.0 and Random Forest. The 161 pancreas MR images are classified as either affected or not affected by using GLCM feature extraction. Among the 3 discrimination methods, C5.0 yields best accuracy (99%) with less error rate.