DVT Assist (Basic, Pro, & Premium)

We deliver the products as a Software as a Service (SaaS) via user-friendly APIs, Apps, or Web interfaces. Subscriptions to the services are available at three level Basic (Free) and Pro/Premium (Paid,) levels. The basic version is designed to provide assistance to primary care professionals and individuals in graphical and voice-enabled common-sense terms (no expertise needed to use, Alex/Siri/Ok google enabled!). The software uses leading edge Artificial Intelligence and Machine Learning (AI/ML). The pro-version is designed for doctors and can use Ultrasound videos, physics based advanced simulations & modeling (CFD) , and data-driven Scientific ML to provide real time (or with little time delay as much as possible) assistance to doctors dealing with serious DVT/PE cases. The premium version is at enterprise level. It is designed to provide detailed highly-resolved results for individual patient from his/her diagnosis/medical data in days using high-fidelity physics based models for approximated geometries obtained from the CT images/Ultrasound videos, pore network models, and HPC/Clould computing resources.

Supporting software developed

• Pore-network-simulation software

Mostly written in C++ using Sandia's scalable and portable libraries (Copyright: The University of Texas System). Developers: Chattopadhyay, Dr. Kotteda, Dr. Kumar, Dr. Delgado, Dr. Harris, and many others from our research group. Thanks to Department of Energy (NETL & Sandia) and Dept. of Defense for support.

• MFIX-Trilinos

C++, Fortran, Python

• Multiphysics AI/ML/UQ Coupling Framework

C++, Python

Source codes: git clone git@bitbucket.org:drkumarresearchgroup/mfix-trilinos.git

Third party software/libraries used

• Trilinos, Dakota, Albany, & Kokkos (Sandia, Opensource), C++

• MFiX (NETL, Opensource), Fortran, Python, C++ (Trilinos integration)

• TensorFlow (Google, Opensource), Python hooks