Repository for master thesis project. Required Python packages specified in requirements.txt, script was written using Python 2.7 version.
prediction pipeline folder contains pre-trained autoencoder for nuclei segmentation and consecutive cell classifier to detect lymphocytes.
The script will classify WSI .svs images stored in slidepath directory, which has to be specified in prediction pipeline/Lymphocyte_prediction_workflow.py line 205.
Predictions will be stored in prediction pipeline/mormin_predictions folder, which will be created at the beginning of prediction workflow.
To execute the code, run the following line in prediction pipeline directory:
python Lymphocyte_prediction_workflow.pySegmentation autoencoder training image augmentation and model training scripts are in autoencoder_training directory;
classifier model was trained using script in classifier_training folder.
Training data is not supplied in this project.