Developing PaaS with Nvidia Clara

Hi,

I was wondering if its possible to use Clara for commercial use by the radiologists.

For my understanding, we first deploy Clara Train on AWS cloud to train models for local data sets. After training the models, we create a clinical workflow where a radiologist interacts with the scan through OHRIF.

Is it possible to create custom DICOM uploader rather than connecting it to a PACS server because in my region, the clinical workflow medical imaging is not up-to-date and manually uploading of DICOM files will be alot easier to handle for the radiologist.

Please let me know if I am understanding Clara the right way?

Thankyou.

Hi talhadar90,

Your general understanding of the Clara Train AIAA workflow is correct. You can learn more about AIAA models provided as part of Clara Train and custom models here:
https://docs.nvidia.com/clara/tlt-mi/aiaa/key_features.html

To interface with the AIAA server, we provide plugins for MITK and 3D Slicer. There is also a reference client implementation on Github. These client options interface with AIAA server and allow for manually uploading local input files. More info here:
https://docs.nvidia.com/clara/tlt-mi/aiaa/client.html

Regarding commercial use, I would defer to the Clara Train SDK EULA, and also highlight the disclaimer on the main Clara Developer page:

Disclaimer: Clara SDKs and samples are for developmental purposes only and cannot be used directly for clinical procedures.

Thanks,
Kris