Feature Classification and Analysis of Acute and Chronic Pancreatitis Using Supervised Machine Learning Algorithm

Supervised Machine learning algorithms play a vital role in prediction of diseases, and the death rate comparatively reduced in twenty-first century; one of the most predominant diseases is acute and chronic Pancreatitis in which only 7% of patients were alive after treatment since the cancerous cells spread to all other parts of the pancreas gland. This marks the importance of diagnosing cancer at an incipient stage and as a result in this work more than 3000 CT scan tumor images from 82 patients around the world are considered. In preprocessing stage the speckle noise is removed using wiener’s filter and then the normalized images are partitioned by enhanced region-based active contour (ERBAC) to find the region of interest (ROI). The features are extracted from the segmented images by gray Llevel co-occurrence matrix (GLCM) and the extracted features are classified by using KNN and SVM classification algorithm. Based upon the results it is found that KNN produces 97.2% of accuracy that the patients are diagnosed at incipient stage. This feature classification may be further improved using artificial neuro-fuzzy inference system (ANFIS) to predict with 99% accuracy.