Brain Tumor Detection From MRI Image Using Matlab

Abstract

One of the most dangerous diesis for the human is a brain tumor, it needs to detect early otherwise nothing remain to do by a doctor or nature. The present situation of detecting tumors depends on the neuro-specialists and radiologists; it is simply that a man or woman can make mistakes or human error can occur. Though the accuracy of manual detection is not bad this process is time-consuming. Magnetic Resonance Imaging (MRI) is the main source for diagnosing the brain tumor. This study describes the way how to find and mark a tumor from an MRI image with high accuracy. In this proposed system the process is too much faster than the manual system, and it doesn’t waste a second if the MRI image has no tumor on it. It can detect if there any tumor exists or not within a few milliseconds and after performing a few analyses if a tumor detected then goes for the further procedure which is time-saving. This thesis also performs the Computer-Aided Detection System (CAD System) which makes it easy to analyze for radiologists and doctors, like how big the tumor, exactly where the tumor is and the shape of the tumor, etc. This thesis is performed by three-stage; the first stage is image pre-processing and postprocessing to enhancement the quality of the MRI image. It makes the image more suitable for further analysis. With a certain threshold label, the main image is converted to a binary image. A statistical property is applied in the second stage to measure all properties of the image region. It extracts the solidity and high-density area from the tumor and performs few operations. In the third stage, the tumor is detected by a few logical operations and make it more visible and mark exactly where the tumor is. The performance was successfully tested and achieved the best result with an accuracy of almost 96%.