Diagnosis of Alzheimer’s Disease Using Convolutional Neural Network with Select Slices by Landmark on Hippocampus In MRI Images

Abstract

Dementia may have numerous origins, the most common of which being Alzheimer's disease. The diagnostic timeframe and an ageing population are projected to lead to a rise in the prevalence of the disorder, which worsens with time and makes it unable to do any work without assistance. Traditional methods of diagnosing Alzheimer's disease are time-consuming and laborious for both patients and clinicians. These methods include neurophysical testing, MRI scans, and retrieval of previous medical information, all of which may be problematic for patients. It is much easier to try to treat brain illnesses if they are detected early. By integrating deep learning with neural networks, our study has enabled earlier-than-usual detection of Alzheimer's disease. Due to the extreme imbalance in the Kaggle dataset, we used SMOTE to distribute the data uniformly throughout the categories. Following this, the model is trained and evaluated using the MRI data that has been classified as very mild, mild, moderate, or severe AD. Afterwards, features are extracted in order to analyse the outcomes. In terms of accuracy and precision, our findings were found to be far better than those of earlier efforts at Alzheimer's identification.