Company Creates High-Fidelity Synthetic Data to Train Artificial Intelligence Models
ASHBURN, Va., July 9, 2020 /PRNewswire/ — Cignal LLC, a technology startup that develops cutting-edge capabilities for the rapid training and deployment of advanced inspection and security systems, announced that it was awarded Phase 1 project funding by the Department of Homeland Security (DHS) Science and Technology Directorate’s (S&T) Silicon Valley Innovation Program (SVIP).
During this project, Cignal will enhance its working prototype and training workflow product, Cignal Workbench, to generate high-fidelity synthetic volumetric data for Computed Tomography (CT) applications and Advanced Technology (AT) X-ray inspection systems. The result will be a virtually unlimited source of labeled training data, leading to automated and continuous training of advanced artificial intelligence (AI) models.
“We are proud to be working with DHS S&T on this very important project,” said Cignal CEO Jaclyn Fiterman. “SVIP’s innovative startup engagement model will allow Cignal to further expand the capabilities and applications of Cignal Workbench and deliver an advanced homeland security solution that provides seamless, unsupervised AI model training on a billion baggage images – a feat impossible today.”
AI models rely on large amounts of labeled training data to learn, and generating this data for screening applications currently is a labor-intensive, manual process. However, Cignal’s creation of synthetic data to train AI models eliminates the need for manual image labeling while increasing the overall amount of quality training data.
About Cignal LLC
Cignal LLC makes software systems for advanced inspection and screening applications. Cignal’s flagship product, Cignal Workbench, helps users generate advanced security and screening models to make transportation and travel safer and more efficient. Cignal LLC is headquartered in Ashburn, Virginia, a hub within the Dulles Technology Corridor. For more information, please visit www.cignal.co.