Prediction of Alzhimers Disease from 4 Class of MRI Images Using Inception V3 Model in Deep Learning

Abstract

Today, Alzheimer's Disease (AD), a degenerative neurological disorder, is the most common cause of dementia. Although it can affect anyone, older folks are often the ones who are affected. Alzheimer's disease (AD) is distinguished by a gradual decline in cognitive function, memory loss, atypical behaviour and personality, and difficulties with daily activities. Although the exact cause of AD is still unknown, a combination of genetic, environmental, and lifestyle factors are thought to be responsible. Several techniques, including machine learning and data analysis, are used in the prediction of Alzheimer's disease (AD), either to identify those who are at risk of contracting the illness or to predict how patients will fair over time. Treatments may be more effective, and patient outcomes may be improved, thanks to early detection and management made possible by AD prediction. There are numerous methods for predicting AD, and they usually include information from the clinical, genetic, and neuroimaging fields. Since the Inception module of the V3 network combines a max-pooling operation with multiple simultaneous convolutional operations with different filter sizes (1x1, 3x3, and 5x5), we attempt to identify AD using the Inception V3 model in this paper. By performing many actions simultaneously, the network can collect properties at different sizes and minimise spatial dimensions, allowing for more efficient representation learning. By running numerous tests on the Inception V3 model using a variety of MRI pictures as input, we eventually outperformed several other models with an accuracy of 77.08%.