NSCLC Deep Learning

This project contains several models to test the prognostic quality of MPM imaging endpoints. While it is specifically designed with NSCLC patient samples in mind, the classes, logical structure, train-test framework, etc. are all easily transferable to other applications. the focus of this project is on determining the best classifier architecture and features to determine treatment response and predict metastatic potential. while classification is the focus, few models are also included to capitalize on the dataset. these models are designed with automated masking, dataset augmentation, and generated redox maps in mind.