Classification of Alzheimer’s and Parkinson’s Disease Based on VGG19 Features with Batch Normalization

Abstract

Dementia is a condition when thinking, reasoning and memory skills are lost and patients have emotional instability and personality changes. Researchers are looking into how the underlying disease processes that lead to various kinds of dementia begin and interact. Additionally, they keep researching the various diseases and conditions that cause dementia. Alzheimer’s and Parkinson's disease contribute to dementia development. Recently deep learning-based techniques have surpassed the performance of traditional algorithms in the field of machine vision, image detection, natural language handling, object detection, and medical image analysis. This study proposed a transfer learning-based model for Parkinson’s and Alzheimer’s disease classification from slices of MRI. Pretrained VGG19 with Batch normalization is used for feature extraction and the final dense (fully connected-FC) layers are fine-tuned to meet our requirements. The performance of the model is analyzed by varying hyperparameters. The proposed model outperformed other pre-trained CNN models by achieving an accuracy of 97.19%.