Early Diagnosis of Alzheimer’s Disease using VGG 19 and RESNET 50

Abstract

Early diagnosis is crucial for Alzheimer's disease (AD), a neurodegenerative condition that affects language, memory, and cognitive function. Complicated language structures and lengthy conversations are difficult for current speech-to-text models to handle. Conventional methods have limitations in terms of accuracy and efficiency, such as analysing MFCC images using CNNs like VGG-16. To address these issues, the suggested system transforms speech into MFCC images and uses sophisticated pre-trained models such as VGG-19 and ResNet50 to analyse them. In order to boost performance, a hybrid model also incorporates these networks with a random classifier. In addition to addressing the drawbacks of speech-to-text systems, this approach minimizes preprocessing steps. The use of image-based analysis to provide quicker and more precise dementia classification helps to improve early detection and diagnosis of Alzheimer's disease.