Proof of concept scripts for running within the secure enclave
step 1 download the repository and unzip it. You have a folder named TREEHOOSE-Enclave-PoC in your local file system;
step 2 create a folder called output within the TREEHOOSE-Enclave-PoC/Main folder;
step 3 download the trained SVM model containing autoencoderMRI.h5 and MRI_Sequence_SVMClassifier.pkl model file into the TREEHOOSE-Enclave-PoC/Main/output folder, the links for the model files are on Teams.
step 4 open the project in your IDE (PyCharm), the project folder is TREEHOOSE-Enclave-PoC/Main
step 5 make sure the following packages are installed for the project
- tensorflow
- sklearn
- opencv-python
- nibabel
- pydicom
step 6 run the script python Main/SVM_classifier_tester.py, you should get a file MRI_Sequence_Results.txt in TREEHOOSE-Enclave-PoC/Main/output folder if you are successful.
Optional
To change the default paths use the script's arguments:
optional arguments:
-h, --help show this help message and exit
-m MODEL, --model MODEL
path to trained autoencoder
-s SVM, --SVM SVM path to SVM index file
-d DATA, --data DATA path image data
-o OUT, --out OUT output path
The defaults are:
python Main/SVM_classifier_tester.py -m Main/output/autoencoderMRI.h5 \\
-s Main/output/MRI_Sequence_SVMClassifier.pkl -d Data/ -o Main/output/