A project that uses convolutional networks to classify the severity of Alzheimer's disease on the basis of images taken magnetic resonance imaging. The program was written in python using the keras and tensorflow packages. The project was made for a Computer Intelligence course at the University of Gdańsk.
Project completed on 8 June 2022
PDF with project report (Polish language)
- keras
- numpy
- sklearn
- tensorflow
Clone down this repository.
Installation:
pip install -r requirements.txt
To Prepare Data:
python prepareData.py
To Run Classification:
python classification.py
- The project allows you to classify the severity of alzheimer's disease.
- Possibility to set options:
- save option - train and save new model
- load option - load ready model from file
- You can set the option to copy wrongly classified images to the wrong_predicted folder.
- You can display statistics:
- train time
- evaluation time of one image
- evaluation time of test set
- accuracy
- error matrix