Comparative Analysis of Multi-disease Prediction using Deep learning
Date
2023
Student ID
Research Supervisor
Editor
Journal Title
Volume
Issue
Journal Title
Journal ISSN
Volume Title
Abstract
This study focuses on the development and deployment of a multi-disease prediction system utilizing deep learning algorithms applied to MRI brain images. The research involves the creation of a specialized network model trained on MRI datasets collected from Kaggle, aimed at predicting various neurological conditions. The proposed model's architecture involves convolutional layers extracting intricate features from the MRI images, transitioning from raw data to pixel-level details, edges, shapes, and finally, disease-relevant regions. Visualization techniques highlight the extracted features and emphasize the network's ability to differentiate between healthy and diseased brain structures. Moreover, the study elucidates the interpretability of the CNN model through weight visualization and modified fully connected layers, delineating distinctions between classes, such as healthy controls and patients with Parkinson's disease. The performance evaluation of multiple deep learning models, including Basic CNN, Modified CNN, AlexNet, and VGG16, is conducted across various epochs. Results showcase notable accuracies in predicting Brain Stroke, Parkinson's disease, and Alzheimer's disease, with each model exhibiting distinct performance metrics. For instance, in Parkinson's disease prediction, transfer learning with VGG16 attains an accuracy of 98.79%, while Basic CNN achieves 91.984% accuracy in Brain Stroke prediction. This comparative analysis highlights the efficiency of deep learning methodologies in diagnosing neurological conditions using MRI data. The findings underscore the potential of these models in facilitating accurate disease identification, paving the way for improved diagnostic tools and interventions in neurology.
Description
Keywords
Citation
Department Name
Electrical and Computer Engineering
Publisher
North South University