Phase Detection Methods for Alzheimer's disease Based on Improved Convolutional Neural Network Fine-tuning

Alzheimer's disease (AD) is a prominent neurodegenerative disease that can cause memory loss and cognitive impairment. Accurate early diagnosis greatly reduces treatment costs and slows progression. Traditional detection methods are complex, expensive, and less accurate. In this study, an Alzheimer's phase detection method based on improved convolution neural network fine-tuning was proposed to diagnose the severity of Alzheimer's disease. The method was trained and tested on Alzheimer's Dataset (4 classes of Images), which classified the patient condition into four stages: MildDemented, VeryMildDemented, NonDemented, and ModerateDemeneted. According to the research results, the classification accuracy on the test set can reach 99.30%. Under the Weighted Average evaluation index, the precision rate of the model is 99.08%, the recall rate is 99.08%, and the F1 value of the model is 99.08%