Page 48 - Security Today, September/October 2023

“... regardless of both the health risk, and that of the risk of a prison sentence, smugglers still see the risk outweighed by the benefit. With today’s prison currency rates, drugs can command up to 10 times higher than current street rates.” As many as 90% of our inmates in Putnam County are incarcer- ated for illegal drug activity whether it be through direct or indi- rect activity,” Sheriff Eddie Farris said. (As reported by WSMV) “Even though inmates are strip searched and scanned while being admitted into our facility, items inside a body cavity are not able to be seen at times,” said Jail Administrator Major Tim Nash. (As reported by WSMV) Removing the risk of operators missing contraband at any lo- cation within the human body is now entirely possible, thanks to Artificial Intelligence and Machine Learning which has now found its way, inevitably into the field of safety and security in the global prison estate. Body scanners are used for contraband detection. These body scanners are made up of an X-ray emitter and an x-ray detector. To perform a scan, an individual stands between the emitter and the detector, and a low dose of X-ray radiation is passed through that person, allowing an X-ray image of that person to be made, detecting both organic and non-organic items. Body scans can find materials such as metal, plastics, ceramics, paper, fluids, flora and even fauna. But as materials get less dense, detection of them becomes less obvious. Netherlands based, ODSecurity has introduced THEIA, a full body automatic threat recognition solution that will detect any type of contraband in any location within the human body. Joost Deutkom, of ODSecurity said, “Real-life operators are es- sential and irreplaceable; But a trained operator with the data, in- sight and accuracy of AI added into the body scanning mix, will in- creasingly become a formidable force in prison safety and security.” Driven by machine learning algorithms, THEIA has been trained extensively on large collections of full-body scan images which are free of contraband. These scans have taught THEIA what a normal human body looks like and, what a negative scan should look like. The contraband free images form datasets, and each dataset teaches the AI technology to recognise anything that deviates from a contraband free full body scan. These deviations, also referred to as anomalies, will be found by THEIA, and highlighted to the human operator. This is what makes THEIA the all-encompassing solution when it comes to automatic contraband detection. THEIA makes no distinction between the detected contraband material, its size, or the location of where it may be hidden. Not only on the human body, but also within it. If it is there, then THEIA will discover it. It is the symbiotic relationship between the human operators and THEIA that have reduced operator error and improved overall safety by direct staff to potential contraband, thereby stopping it from entering our prison and custodial estates. An enormous benefit to using the machine learning and AI algorithms that make up THEIA, is the software will continu- ously improve. The more data THEIA is exposed to, the more it learns, optimizes, and grows in accuracy – this in many ways, “future proofs” contraband detection, as unusual, or contraband that has not yet been attempted, discovered, or indeed created will be found by the simple exclusion of anomalies. Narcotics come in many shapes, forms and substances. Some are relatively easy to detect, others harder to distinguish from the human body, as their organic nature disguises them, for example, large quantities of drug balloons in intestines, are relatively easy for operators to detect through normal scanning, due to the den- sity of the package and the shape, although smaller quantities can be more difficult to find. Also, organic materials like canna- bis, and pills in a body cavity can be a challenge for operators, as they too can blend in with the human body. But, with the use of AI, these items are not perceived as being naturally part of the human body and are then highlighted for an operator to review. As we all know, it is not just contraband drugs that cause mas- sive problems in our prison estates. Everyday communication and electronic devices are getting smaller. This makes them easier to hide and harder to detect. Thumb drives can be stripped down to the bare minimum leaving a tiny device with minimal amounts of metal, and SIM cards are so small they are being hidden in body cavities to be brought into prisons. All these items can be easily overlooked when concealed in the right spot. But then THEIA is designed to find those items that are not naturally part of a hu- man body – and will highlight them for review. Weapons in a prison environment also become a valuable commodity. Bullets are easily detected by a body scanner but can be missed by a metal detector. Razor blades could be seen if they were perpendicular, but not if they were in line with the detector, as you would just see a very thin line. Then again, a ceramic knife with a bone handle, could be picked up by a detector, but possibly not if it was placed close to the rib cage or spine. But with THEIA, all the above would be found. As they are not naturally part of the human body, and therefore be flagged for review. Better safeguards to stop the influx of contraband into our prisons is not just a wish, but a necessity, and whilst the human body scanner operator will be successful at stopping some of the contraband, as he naturally visually searches for suspicious anomalies, in or on a human body, joined by the AI assistant, that is looking for what “shouldn’t be there” will prove a powerful weapon in the contraband de- tection arsenal. Karen Kingham is a journalist and PR special- ist who has been working in the security and defense sectors for more than 25 years. 48 SEPTEMBER/OCTOBER 2023 | SECURITY TODAY ARTIFICIAL INTELLIGENCE