Police could identify paedophiles online using AI hand recognition

Researchers are asking for help from the public to help build a database of hands that would allow police to identify tens of thousands of paedophiles every day.

Forensic scientists currently can link suspects to child abuse footage, through analysing the back of the hands in the footage and whether things like blood vessels map up to those on the hands of a suspected child offender.

However, the process is thought to be very slow, with one case taking at least two weeks, and the success rate at around 86pc.

Researchers from Lancaster University and the University of Dundee are attempting to automate this through building a hand recognition system. This will mean that police should be able to find a suspect for the child abuse, without having had to identify a perpetrator first. This system should be able to analyse tens of thousands of images from such cases every day.

Bryan Williams, the lead researcher on the H-Unique programme, said it would also make it easier to use the evidence in court cases, improving the accuracy significantly and “strengthening the cases we’re doing now”.

To build these algorithms, though, Mr Williams said the scientists needed a “huge variety of hands”.

“We want to be able to identify how to pick out blood vessels on white skin, on Chinese skin, so we need a huge variety of ethnicities and a huge variety of age ranges as well,” he said.

“We’re asking the public to use our app to take pictures on their phones, and then send them into the project.”

The images will be submitted anonymously, will not be shared with external agencies, and will be destroyed at the end of the project.

The programme is aiming to get at least 5,000 images over the next three years, although it is hoping to collect many more images and in a much quicker time-frame.

According to research from the Crime Survey for England and Wales, released earlier this year, around one in five adults had experienced some form of child abuse before turning 16, ranging from threats to rape.