tanyajoon / parkinson-disease-detection-using-pet-images

Parkinson Disease Detection using PET Images

Execution Flow

Dataset preparation

1. Acquire data set (*.nii files) from PPMI
2. Acquire the label files (*.csv file)
3. Update the 'dataPath' variable (if placed somewhere else)
4. Run the dataset preparation file using python3 and Jupyter Notebook.

Dataset preprocessing

FSL is used for preprocessing purposes, for a guide on how to install and use FSL please refer to installation guide. After successful installation, execute the files using python3 and Jupyter Notebook. The files(*.nii) can be visualized using FSL eyes.

Testing and Training:

Before executing the testing and training steps please refer to installation guide and install the required modules. The 3 models (SVM, RF, CNN ) can be executed using the files provided under the Models folder.

CNN

After aquiring the data (data.csv) open Model_CNN.ipnyb adn Model_CNN Kfold file found in tghe Model folder.

SVM

After aquiring the related data (Data.csv) open Model_SVM.ipnyb file found in tghe Model folder.

RF

After aquiring the related data (Datacnn.csv) open Model_random forest.ipnyb file found in tghe Model folder.

OUTPUT

after running all of the above mentioned scripts, you will have a final average accuracy at the end which will be that models accuracy without cross validation, cross validated result is calculated and shown before the main model in SVM, RF and a different file (Model_Cnn Kfold.ipnyb) contains cross validation for the CNN model.

Demo Video:

A video guide on how this project works and some general guidelines are available in the following youtube link: Youtube