This Jupyter notebook focuses on the application of Convolutional Neural Networks (CNN) for detecting Alzheimer's disease from brain scans. Utilizing advanced machine learning techniques, specifically Xception and EfficientNetB7 models, we aim to classify brain images accurately, identifying markers of Alzheimer's and differentiating between the stages of the disease. This approach, leveraging transfer learning, seeks to enhance diagnostic precision with the limited data available.
The brain scan images used in this study are sourced from the following Kaggle dataset: Alzheimer's MRI Brain Scan Images Augmented.
The notebook and related materials are available on GitHub: DTSA-5511 Introduction to Deep Learning Final Project.
- Download Data: Follow instructions to download the dataset from the above Kaggle link.
- Load Dataset: Instructions and code for loading the dataset into dataframes for analysis are provided in the notebook.
- Model Training and Evaluation: The notebook includes comprehensive steps for training the Xception and EfficientNetB7 models, followed by an evaluation of their performance.
- Prediction and Analysis: The final sections of the notebook are dedicated to making predictions with the trained models and analyzing their performance.
This notebook is designed to run in a Python environment with specific dependencies, including TensorFlow, Keras, NumPy, and Pandas. For a complete list of requirements, please refer to the included requirements.txt file in the GitHub repository.
Contributions to the project are welcome. Please refer to the GitHub repository for contributing guidelines.
This project is open-sourced under the MIT License.