The inside of a bag viewed with UCL’s X-ray system. (Photo: Patridge et al/Nature Communications)A new X-ray technique that combines conventional equipment with a deep-learning algorithm might find its way into both security settings and the healthcare industry.
Researchers from the United Kingdom’s University College London (UCL) recognized that X-ray security systems, though good at detecting shapes, weren’t so great at recognizing textures. Identifying textural abnormalities could be the key to locating explosives and other harmful items—especially those hidden away within larger objects. So they set about devising a system that could be paired with existing equipment to detect concerning textures.
Physically, this consisted of “masks,” or sheets of metal with tiny holes punched through them. These served to separate X-ray beams into smaller beamlets, which scattered at incredibly tight angles (as small as a microradian, or about one 20,000th of a degree). This resulted in a more defined image: a flat, dark image became sharp and nearly three-dimensional. The researchers then created an algorithm that could analyze the scattering. These scatter patterns would be used to recognize the textures of different materials.
Once the deep-learning algorithm had been trained, the X-ray machine’s level of accuracy was astounding. The researchers placed explosives like C4 and Semtex into travel bag replicas containing toothbrushes, phone chargers, and other basic objects. They also tucked the explosives into devices like laptops, phones, and hair dryers. Thanks to the system’s ability to detect textures within other textures, it was able to detect the explosives with 100 percent accuracy, according to a paper published last week in Nature Communications. Dangerous substances were visible under the new system’s watchful eye, while those same substances quite literally slipped under the old system’s radar.
Airports aren’t the only setting that might benefit from this system. Though the researchers haven’t yet tested the rigged X-ray’s ability to detect tumors within healthy tissue, they believe the system might someday be useful for early cancer screening. One such use case might involve detecting small tumors hidden behind a person’s ribs. Architects might even be able to alter the system to detect rust and cracks in buildings.
Given the X-ray’s 100 percent success rate under test conditions, the UCL researchers are interested in trying out their system in real-life settings—or at least lab environments that imitate real life more closely.