The USA’s Department of Homeland Security (DHS) Science and Technology Directorate (S&T) has launched a dataset of 250 images of synthetic carry-on baggage to improve screening operations.
The new dataset provides original equipment manufacturers (OEMs) and third-party AI developers with access to high-fidelity, physically accurate x-ray and CT synthetic images. This resource is designed to help developers build, test and validate automatic threat recognition algorithms against complex scenarios without the bottlenecks of physical data collection.
“Detection algorithms are key to reducing false alarms that cause hands-on bag inspections and longer checkpoint wait times,” said Pedro Allende, DHS Under Secretary for Science and Technology. “Enabling rapid algorithm development from the best innovators not only will increase detection accuracy at the checkpoint but also improve the traveler experience.”
“Currently, only a limited number of images of curated bags are available to train carry-on baggage algorithms, and developers must complete a vetting process to view them, which is prohibitive for some,” said DHS’s Screening at Speed program manager, Dr John Fortune.
To support S&T, Cignal has developed synthetic images for the Screening at Speed program. The dataset, available by request on Cignal’s website, includes 250 computed tomography (CT) x-ray images that mirror the high-resolution scans produced by the Transportation Security Administration’s latest detection systems. The images will be released in Digital Imaging and Communication in Medicine (DICOM) format, familiar to algorithm developers across industries.
S&T funded the development of the Cignal Engine technology through the Silicon Valley Innovation Program. Future releases may feature larger, more complex bag sets and synthetic millimeter-wave images for training passenger screening systems.
In related news, Poczta Polska wins new three-year security contract at Gdańsk Airport
