Brain Tumor Detection Using Image Processing and Convolutional Neural Network MATLAB

Abstract

Even with the tremendous advancements in medical technology, detecting brain tumours is still a time-consuming and difficult endeavour for doctors. Brain tumours may be treated more effectively and efficiently if they are detected early and accurately. An improved level of predictability in the diagnosis and categorization of brain tumours may be achieved by using automated systems. As a result, it is well documented that the accuracy performance of automated detection and classification approaches varies from technique to technique and tends to be reliant on the picture modality used. This study examines the current detection methods and outlines their advantages and disadvantages. As one of the most frequent illnesses in the world, brain tumours may be described as the unchecked growth of abnormal cells in the brain, which presents a diagnostic difficulty. When paired with a well-established image processing technology, identification of this condition becomes much simpler. The goal of this research is to propose a feasible approach for fast determining the size and location of a tumour from an MRI image utilising area splitting, merging, and growth based segmentation. A total of five phases are involved in the whole process, including input in the form of MRI pictures, pre-processing and enhancement, image segmentation and feature extraction. In order to identify the tumour, MRI images were enhanced using contrast enhancement and median filtering, which was followed by a segmentation technique. Graphical user interfaces and MATLAB algorithms have been used to organise input and output data.

Keywords: MRI, GUI, Segmentation, Brain Tumor, Filtering, Enhancement, MATLAB.