NeurodegenerationClassification
Postgraduate assignment using Neural Networks to classify Neurodegeneration
Abstract
Neurological degeneration is a significant and irreversible medical condition where the neuron structure within the brain is destroyed. Conditions such as Alzheimer’s Disease, are slow processes which can cause issues with the patient’s memory, social function and personal behaviour. Early detection of Alzheimer’s Disease can allow for some intervention treatments to mitigate the impact. This paper proposes some novel detection methods for dementia through the use of convolutional neural networks (CNN) of MRI brain images. Some different CNN structures are presented in order to improve the classification performance. The OASIS dataset of 373 MRI images of dementia patients is used for training and classifying the CNNs. The ensemble method combining multiple CNNs had the best validation accuracy of 72%. This automatic method demonstrates viability and would supplement professional medical judgement in the early detection of Alzheimer’s Disease and other similar conditions.
Report
Dataset
Scripts
Scripts
Data
Images
