Swallowed Drugs X-ray Screening

Criminal gangs have made millions from within prison cells, selling heroin and spice to the outside world using cell phones the size of domino pieces.

Inmates have been able to swallow or leak lethal drugs undetected because prison officers are not authorized to see into their bodies. They then wreak havoc within jails, with some inmates overdosing, forcing guards to dispatch ambulances that could have been spent elsewhere.

The Prison Officers Association had requested the government to install X-ray body scanners in all closed men's prisons in England and Wales, and they have already been deployed. They have foiled over 20,000 attempts to smuggle in harmful and life-threatening chemicals in just over a year.

The scanners are so successful in disrupting organized crime groups (OCGs) that dealers have offered a £10,000 reward to anyone who can damage the equipment sufficiently to disable it, according to governors.

Hundreds of inmates have already been captured attempting to enter the facility with narcotics, tobacco, and other illicit materials.

Based on the latest research, at least 12,500 ambulances and paramedic cars were dispatched to English jails in 2019 and 2020. Over 200 people died as a result of an overdose or a stabbing. According to the most recent data, every ambulance call that resulted in a patient being transported to hospital cost NHS England £292 in 2019/20.

The method of automatic swallowed drugs detection

• A method for an automated detection of swallowed capsules on X-Ray scanner images, the method comprising:
• Acquiring an incoming image of a person passing through the scanner;
• Generating additional images based on the incoming image by performing transformations of the incoming image;
• Determining an upper body area on the incoming image;
• Calculating a position of a stomach and genitals areas on the upper body area of the incoming image and on the additional images;
• Classifying segmented regions of the stomach and genital area of the incoming image;
• Calculating geometrical and intensity features along with rotationally invariant periodic features for windows on the stomach and genital areas;
• Detecting suspected windows on the stomach and genital areas;
• Calculating aggregate features for the properties of the suspected windows;
• Using a model for the images that do not contain swallowed capsules classifying the incoming image using a threshold for a dissimilarity function;
• Informing operator that the incoming image contains the swallowed capsules if the dissimilarity function for aggregate features for the incoming image is larger than or equal to the threshold.