This post is also available in: עברית (Hebrew)
Checking travelers property at airports demands skilled staff. A startup will develop a proof-of-concept for an artificial intelligence (AI)-based object recognition capability for the US Transportation Security Administration (TSA), that could greatly enhance security checkpoint operations.
The Department of Homeland Security (DHS) Science and Technology Directorate (S&T) awarded $199,961.29 to Synthetik Applied Technologies, for a solution that “proposes an advancement to three-dimensional (3D) object recognition” said Melissa Oh, S&T’s Silicon Valley Innovation Program (SVIP) Managing Director. “The solution has many applications to DHS use-cases that we have not yet fully explored.”
The focus of this development is a capability using real-time voxel-wise instance segmentation to detect objects during property screening in airports.
A voxel is a unit of measurement in a 3D image—a 3D pixel—and object segmentation is the problem of delineating each object of interest in an image.
Voxel-wise instance segmentation uses AI to identify objects within a 3D image. The solution proposes to develop and train an AI model with the potential to automatically detect multiple objects at the same time during the property screening process at an airport, enhancing current human-based capabilities, according to newswise.com.
The company is currently leading research with the U.S. Defense Advanced Research Projects Agency (DARPA) to fundamentally redefine how to process 3-D data using artificial intelligence. It is also supporting the U.S. Department of Homeland Security (DHS) by employing computer vision and deep-learning methods for automatic anomaly detection at speed, as well as working directly with NOAA and Microsoft AI for Earth to develop a low-cost entanglement mitigation system to protect endangered marine species.
SVIP is one of S&T’s programs and tools to fund innovation and work with private sector partners to advance homeland security solutions.