Pareto Optimized Adaptive Learning with Transposed Convolution for Image Fusion Alzheimer’s Disease Classification

• DOI: 10.3390/brainsci13071045
• Publication Date: 2023-07-10T04:35:21Z
• Authors: Modupe Odusami; Rytis Maskeliūnas; Robertas Damaševi...
• Keywords: Feature (linguistics)

Alzheimer's disease (AD) is a neurological condition that gradually weakens the brain and impairs cognition memory. Multimodal imaging techniques have become increasingly important in diagnosis of AD because they can help monitor progression over time by providing more complete picture changes occur AD. Medical image fusion crucial it combines data from various modalities into single, better-understood output. The present study explores feasibility employing Pareto optimized deep learning methodologies to integrate Magnetic Resonance Imaging (MRI) Positron Emission Tomography (PET) images through utilization pre-existing models, namely Visual Geometry Group (VGG) 11, VGG16, VGG19 architectures. Morphological operations are carried out on MRI PET using Analyze 14.0 software after which manipulated for desired angle alignment with GNU Image Manipulation Program (GIMP). To enhance network's performance, transposed convolution layer incorporated previously extracted feature maps before fusion. This process generates weights facilitate process. investigation concerns assessment efficacy three VGG models capturing significant features data. hyperparameters tuned optimization. models' performance evaluated ADNI dataset utilizing Structure Similarity Index Method (SSIM), Peak Signal-to-Noise Ratio (PSNR), Mean-Square Error (MSE), Entropy (E). Experimental results show outperforms VGG16 VGG11 an average 0.668, 0.802, 0.664 SSIM CN, AD, MCI stages (MRI modality) respectively. Likewise, 0.669, 0.815, 0.660 (PET