E2FP-Net: Elevated Fusion Feature Pyramid Network for Alzheimer Detection

Alzheimer's Disease (AD), a neurogenerative disorder which is a progressive disease and it affects the cognitive function of the brain and aids in behavioral changes and loss of memory. AD a challenging and a complicated condition in which enhanced diagnosing method, treating method is investigated by the researchers. Treating the symptoms of AD can improve the quality of life to the individuals. Diagnosing methods of AD may include Neuro imaging, CSF analysis, Cognitive testing, Genetic testing etc.., By integrating the MRI images and genetic images we propose an Elevated feature pyramid network based on Adaptive hypergraph neural network. Furthermore, the GNN involves in the process of feature extraction of MRI images and Genomic data. Our proposed framework has the following process for detecting AD: A) Twin data pre-processing B) E2FP based Feature extraction C) Node mingle fusion. The feature extraction process has 3 subprocess includes a) Region based feature extraction for MRI images using Adaptive HGNN b) Mid-Level Feature Extraction for Genomic data using Adaptive HGNN c) Amalgamation of Global Feature. Our model performs well in terms of accuracy, sensitivity and specificity. This model has the achieved a great interpretability in genomic data and by enhancing the contrast the MRI data shows a best accuracy.