This document lists the execution flow and the list of files in the codebase
Training+Testing
- Acquire the preprocessed training and testing data set (*.npy files),
- Acquire the label files (*.csv file)
- Place all the files in 'data' folder
- Update the 'dataPath' variable (if placed somewhere else) and the saveModel flag (if you plan to save the trained model)
- Run main.py file using python3.
Only Testing
- Acquire the preprocessed testing data set (*.npy files),
- Acquire the label files (*.csv file)
- Acquire the saved model file (*.pkl files),
- Place the data and label files in 'data' folder and model files in 'models' folder,
- Update the 'dataPath' variable and the model file names/location (if placed somewhere else)
- Run demo.py file using python3.
- preprocessing/ - Contains the shell scripts to preprocess the images.
- models/ - location of saved SVM models
- data/ - location of training and testing data
- dataAcquisition/ - random scripts
- createCSVDataset.py - create train.csv and test.csv to store the location and names of all related files for a subject along with the class
- csvToDat.py - create *.dat files from *.csv file for use with WEKA
- main.py - The main code file. Run this to train and test the model.
- demo.py - Reads the saved model file and predicts the classes for the test dataset
- saveDataNp.py - Methods to read and save train and test data as *npy files.
Training/Testing
- pandas
- pickle
- scikit-learn
- numpy
- nibabel
For Data Preprocessing
- dcm2niix or mricrogl_lx
- fsl
- eddy
- bet
- dtifit
Code available at - https://github.com/HarshSharma12/dti-pd-svm Demo video is available at - https://drive.google.com/a/ualberta.ca/file/d/1t0ADAqo9lDGTKzxX5tc2LduHRVa8eLCq/view?usp=sharing