utilities functions for QARC related project

Refer here https://dicom.nema.org/Dicom/Geninfo/brochure/RTAAPM.HTM

All work done so far is available in this repo https://github.com/kundan1974/QARC-E001

Check out this for preprocessing overview https://github.com/kundan1974/QARC-E001/blob/master/README.md#image-preprocessing--labels

All python module dependency has been cpatured in requirements.txt. You can use install_deps.sh shell script. Alternatively run following command to get all Python modules installed in one go pip install -r requirements.txt

• Python script preprocess.py contains logic to preprocess all patient images. For every patient, once the numpy file has been successfully created, the original patient folder is moved to a specified completed folder. If any error is encountered during preprocessing than that patient folder is moved to an error folder under the specified completed folder In the overall workflow, DICOM images will be fetched from server and numpy files will be split in test and train sets for CNN model to use as inputs
• Shell script preprocess.sh runs python script preprocess.py. Here you can edit input/output folders to suite your environments
• setup_preprocess.sh bash scripts adds a crontab entry to schedule running preprocess.sh at 22:00 hours every day. This will ensure any new DICOM patient images added will be automatically processed on next run

Models based on Cox Proportional Hazard Model and Kaplan Meir Analysis https://github.com/thoraciclang/Deep_Lung

https://ieeexplore.ieee.org/document/9661330/ - IEEE Paper Presentation - https://youtu.be/u4H3KtGxM7I