AI-based Alzheimer's Disease Detection using Convolutional Neural Networks

This repository contains code and resources related to the development of machine learning models for the detection of Alzheimer's disease using Convolutional Neural Networks (CNNs).

Requirements

• Python 3.x
• Tensorflow pip install tensorflow
• Keras pip install keras
• Seaborn pip install seaborn
• Scikit-learn pip install scikit-learn
• Scikit-image pip install scikit-image
• Pydot pip install pydot
• Graphviz here

For GPU Training:

Follow this Tutorial: here

• CUDA Toolkit here

Dataset here

About Dataset

Alzheimer MRI Preprocessed Dataset (128 x 128)

• The Data is collected from several websites/hospitals/public repositories.
• The Dataset consists of Preprocessed MRI (Magnetic Resonance Imaging) Images.
• All the images are resized into 128 x 128 pixels. -The Dataset has four classes of images.
• The Dataset is consists of total 6400 MRI images.
  1. Class - 1: Mild Demented (896 images)
  2. Class - 2: Moderate Demented (64 images)
  3. Class - 3: Non Demented (3200 images)
  4. Class - 4: Very Mild Demented (2240 images)

Motive

The main motive behind sharing this dataset is to design/develop an accurate framework or architecture for the classification of Alzheimers Disease.

References

• https://adni.loni.usc.edu/
• https://www.alzheimers.net/
• https://www.kaggle.com/datasets/jboysen/mri-and-alzheimers
• https://ieeexplore.ieee.org/document/9521165
• https://catalog.data.gov/dataset/alzheimers-disease-and-healthy-aging-data
• https://www.nature.com/articles/s41598-020-79243-9
• https://cordis.europa.eu/article/id/429468-the-final-epad-dataset-is-now-available-on-the-alzheimer-s-disease-workbench

Code

The code in this repository includes data preprocessing, image augmentation, model creation and evaluation using CNNs, and transfer learning techniques. The CNN models are implemented using the Keras API in Tensorflow. Transfer learning is performed using pre-trained CNN models such as VGG and ResNet.

Results

The results of the experiments conducted using CNN models are presented in the form of accuracy and loss as shown below:

• Validation loss: 0.0687 or 6.87%
• Validation accuracy: 0.9805 or 98.05%
• Number of misclassified images: 25 of 1280

Conclusion

This repository provides a comprehensive framework for the development and evaluation of machine learning models for Alzheimer's disease detection using CNNs. The code and resources in this repository can be used as a starting point for further research and development in this area.