The U.S. Department of Homeland Security (DHS) has released a new synthetic X-ray image dataset designed to improve artificial intelligence algorithms used in Transportation Security Administration (TSA) baggage screening, according to media reports.

Developed by Cignal under the DHS Science and Technology Directorate’s Silicon Valley Innovation Program (SVIP), the dataset contains 250 synthetic X-ray images that closely replicate scans produced by TSA detection systems.

The resource is intended to help Original Equipment Manufacturers and third-party AI developers train, test and evaluate threat detection algorithms without relying on sensitive operational data.

The initiative aims to reduce false alarms, improve detection accuracy and ultimately enhance passenger experience by shortening security checkpoint wait times.

“The AI systems protecting our checkpoints require rigorous, dynamic environments. By providing developers with access to high-quality synthetic scenes, we are ensuring that highly adaptive threat detection models can be deployed faster, securely, and with greater accuracy,” said Jaclyn Fiterman, Cignal CEO, in a statement.

The release complements DHS’s broader work under SVIP, which has previously supported AI models for object identification and multi-object detection in airport screening systems.
The announcement also aligns with TSA’s Gold+ initiative, a public-private partnership model that encourages airports to adopt advanced screening technologies and operational services through private-sector collaboration. Participation in the program is voluntary, with several airports already opting in.