Paper Title
Deep Learning Approach for Brain Disease Diagnosis

Abstract
Brain tumour is an unusual mass of tissue in which some cells multiply and grow uncontrollably. Early brain tumour diagnosis plays a crucial role in treatment planning and patients' survival rate. Manual brain tumour detection is complicated, time-consuming, and vulnerable to error. Hence, automated computer-assisted diagnosis at high precision is currently in demand. It needs to be detected at an early stage using MRI or CT scanned images when it is as small as possible because the tumour can possibly result in cancer. This paper mainly focuses on detecting and localizing the tumour region existing in the brain by proposed methodology using patient's MRI images with the help of MATLAB. To pave the way for morphological operation on MRI image, the image is first filtered using Anisotropic Filter. It helps to reduce contrast between consecutive pixels. Anisotropic filtering is a method of enhancing the image quality of textures on surfaces of computer graphics that are at oblique (slant) viewing angles with respect to the camera. It is superior to many other filtering methods as it has a high PSNR (Peak signal to noise ratio) and low MSE (Mean Squared Error). After that, the image will be resized, utilizing a threshold value. The image is then converted to a black and white image. This is done to perform processing techniques on it. Classification of MRI images is an important part to differentiate between normal patients and a patient who has a tumour in the brain. For this, Support Vector Machine (SVM) classifier is used. After the SVM classifier, morphological operations will be applied to obtain information on areas of the possible tumour locations. Morphological operations include dilation and erosion. Dilation is making the objects more visible and Erosion is making objects less visible. Using both dilation and erosion, the tumour outline can be obtained. Then it is used to deliver final detection result, i.e., detect and isolate the region of tumour. This method can be employed to accurately locate the tumour when there is large data available. Using this can also help avoid human error during manual detection of tumour. Keywords - MRI, CT Scan, Anisotropic, SVM, Morphological.