Researchers from the University College London have developed a method for pairing deep learning artificial intelligence (AI) and X-ray machines to detect explosives in luggage.

Following earlier research that determined X-rays, when they hit specific materials, will produce tiny bends that will vary according to type of material, the researchers modified an X-ray machine by adding a box with masks, which are metal sheets featuring tiny holes.

Schematic of edge-illumination x-ray imaging. This is shown in panel a, with a zoom-up on the region between the two x-ray masks in panel b (without object). The x-ray beam is split into a plurality of beamlets by a pre-sample mask (M1). These are then interrogated by a second, analyser mask (M2) placed before the detector, which allows assessing their reduction in intensity (attenuation signal), lateral deflection (refraction signal), broadening (dark-field signal). Source: Nature Communications (2022). DOI: 10.1038/s41467-022-32402-0Schematic of edge-illumination x-ray imaging. This is shown in panel a, with a zoom-up on the region between the two x-ray masks in panel b (without object). The x-ray beam is split into a plurality of beamlets by a pre-sample mask (M1). These are then interrogated by a second, analyser mask (M2) placed before the detector, which allows assessing their reduction in intensity (attenuation signal), lateral deflection (refraction signal), broadening (dark-field signal). Source: Nature Communications (2022). DOI: 10.1038/s41467-022-32402-0

According to the researchers, the masks split the X-ray beam into several smaller beams. The device was then used to scan objects featuring embedded explosive materials and those results were then fed into a deep learning AI application wherein the machine was taught what tiny bends corresponded to what materials.

Once the machine was trained, researchers then scanned another set of objects featuring embedded explosives and determined that the machine could identify the explosives with 100% accuracy in the lab.

In addition to eyeing the technology for transportation security applications, the researchers are exploring if it can also be used in healthcare applications, for identifying a tumor, for instance, as well as for building and airplane inspection applications.

The study, Enhanced detection of threat materials by dark-field x-ray imaging combined with deep neural networks appears in the journal Nature Communications.

To contact the author of this article, email mdonlon@globalspec.com