Since Alzheimer's disease is a neurological condition that affects millions of people worldwide, early detection is vital for enabling successful treatment. Recent advances in deep learning, as demonstrated by models like InceptionV3, show significant potential for precisely forecasting Alzheimer's disease by examining neuroimaging data. This study makes use of a sizable dataset of magnetic resonance imaging (MRI) scans, which includes both participants with Alzheimer's disease (AD) and those who are cognitively normal (NC). The work uses transfer learning to improve an InceptionV3 model that has already been trained for multi-class classification. Brightness adjustment is a component of data preparation, which increases the size of the dataset and strengthens the generalization abilities of the model. In order to overcome concerns with class imbalance in the dataset, pattern classification methods like SMOTE (Synthetic Minority Over-sampling Technique) are also strategically utilized. Following these preprocessing steps, the dataset is divided into training, validation, and test sets to facilitate model training and evaluation. The proposed system produces a good accuracy of 87.69% for the OASIS dataset.
Revolutionizing Alzheimer's Disease Prediction using InceptionV3 in Deep Learning
22.11.2023
359561 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch