CONTENT PREVIEW
Security

AI software could improve CT scanners

18 December 2018
Screenshot of a JSS-01K scan, highlighting the different types of electronic equipment detected. Source: JLK Inspection

Artifical intelligence (AI)-based x-ray inspection software from South Korea-based JLK Inspection is being adapted to work with computed tomography (CT) scanners for checkpoint screening.

The JSS-01K can detect electronic items such as laptops, smartphones, hard disks, USB sticks, and tablets even if one of these items overlays another, said JLK client executive Chang Hoon Kang. As cabin baggage passes through the x-ray tunnel, an analysis takes place in around 0.1 seconds. Target items are highlighted on a monitor display in red (laptops), yellow (hard disks), green (USBs), and so on, obviating the need for any interpretation by operators.

Demonstrations of the system, seen by Jane’s at the International Security Expo in London in November, involved a Rapiscan 620 SR system, although Kang said that the JSS-01K is compatible with all kinds of x-ray machines and it is being adapted to function with CT scanners.

JLK launched the JSS-01K earlier in 2018, having previously collaborated with the Korean Customs Service, which provided an initial database of 40 million images. This allowed the JLK to build its own database by studying each image with AI-based machine learning technology. The company also offers the JPS-01K, which can locate guns, knives and explosives; and it is developing the JAX-01K as its newest security solution for the airport.

“Although similar technologies exist in the market, none are as accurate as ours,” claimed Kang. However, accuracy is affected by the sensitivity level set by the AI engine. Under normal settings, around 90% of smartphones and hard disks are identified, falling to 75% for USB sticks (given the varied types of stick that exist).

By changing the sensitivity parameters with a single mouse click in the user interface, these percentages will rise, but so will the number of false alarms.

Want to read more? For analysis on this article and access to all our insight content, please enquire about our subscription options at ihs.com/contact



(317 of 386 words)
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT