ALZHEIMER DISEASE DETECTION BYCOMPUTER AIDED BRAIN DIAGNOSIS USING BRAIN MRI IMAGES BASED ON DECOMPOSITION TECHNIQUE SAND CONVOLUTIONAL NEURAL NETWORK

Abstract

The main goal of this thesis is to create a system which can determine whether the patient is Alzheimer affected or not and that system we call it as a CABD (System Computer Aided Brain Diagnosis). T2 weighted MRI have been taken as an input. In this thesis, series of quantitative techniques have been taken such as filtering, feature extraction and KNN classifier. The motive of this research work is to develop an interface which can classify normal and Alzheimer Disease cases. Firstly, we collect the 2D test images of Brain MRI. Now after collecting these 2D test images, we implement a preprocessing technique on these images, the result of which is the enhancement of test pictures and removal of the noise. The preprocessing employs a median filtering algorithm which automatically separate noise from the test images. Then from those median filtered Brain MRI, features were collected using various transforms which are describe further. Now, at the final stage we implement KNN classifier which then classifies that whether the person is Alzheimer affected or not.