Tools
The tools are separate python3 script located in the tools folder.
- dataset_statistics.py
Get some stats about dataset.
Example on the Bullitt_Isotrope_light_mult32 dataset :
GLOBAL IN MAX : 1270
GLOBAL IN MIN : 0
GLOBAL IN AVG : 48.05026120061352
GLOBAL GD MAX : 1
GLOBAL GD MIN : 0
GLOBAL GD AVG : 0.004222146971232689
- evaluator.py
Evaluate a prediction Example with a trained model :
Sørensen–Dice coefficient with Prediction 0.6555862420491801
Sørensen–Dice coefficient with Raw Input 0.028533736504787497
Average precision score with Prediction 0.47484733206912355
Average precision score with Raw Input 0.24825881023935145
-
isotrope.py
See more information on Preprocessing page. -
log_analyzer.py
Analyse training.log file to see the metrics with graphics.
Example on a training :
-
mha_to_nii.py
Convert mha file to nii file. -
model_summary.py Load and print a model architecture information using the summary() method from keras model.
-
prediction_mixer.py
Mix predictions to create a new one (operations : max/min/avg) -
shrink.py
Resize NIFTI files with "max pooling" method