EARLY DETECTION OF ALZHEIMER’S USING DEEP LEARNING

Computer Engineering

There are many different causes of dementia, but Alzheimer's Disease is the most usual form. As the condition progresses, it limits one's ability to perform any task without aid, and the diagnosis timeline and aging population are expected to cause its prevalence to increase. The conventional ways of detecting Alzheimer’s is tiring for both patients, doctors where it involves retrieving the past medical records and having Magnetic Resonance Imaging scans and even neurophysical testing which can be inconvenient for patients. An early diagnosis of brain diseases makes a big difference when it comes to attempting to cure them. Our work has used deep learning (neural networks) to detect Alzheimer's disease earlier than usual by combining it with deep learning. As the obtained dataset from Kaggle is heavily imbalanced, we evenly distributed the data between the categories using SMOTE. Then the model is trained and tested with the categorized MRI data i.e. very mild, mild, moderate and severe AD and finally extract features to examine the results. The results we achieved are compared with the previous attempts on detection of Alzheimer’s and came out to be significantly greater in terms of precision and accuracy.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.