Early detection and proper treatment of brain tumors are imperative to prevent permanent damage to the brain even patient death. The present study proposed an AI-based computer-aided diagnosis (CAD) system that refers to the process of automated contrast enhancement followed by identifying the region of interest (ROI) and then classify ROI into benign/malignant classes using significant morphological feature selection. This tool automates the detection procedure and also reduces the manual efforts required in widespread screening of brain MRI. Simple power law transformation technique based on different performance metrics is used to automate the contrast enhancement procedure. Finally, benignancy/malignancy of brain tumor is examined by neural network classifier and its performance is assessed by well-known receiver operating characteristic method. The result of the proposed method is enterprising with very low computational time and accuracy of 87.8%. Hence, the proposed method of CAD procedure may encourage the medical practitioners to get alternative opinion.
Computer-Aided Diagnosis of Low Grade Endometrial Stromal Sarcoma (LGESS)
