As It Happens

How photos of your hands could help bring child predators to justice

A team of forensic researchers in the U.K. are trying to develop an algorithm to catch child predators — and they need thousands of photographs of people's hands to do it.

Forensic anthropologists developing algorithm to identify perpetrators' hands from child pornography

Researchers at Lancaster University and the University of Dundee are developing an algorithm that can scan images of hands for identifying information like skin creases and vein patterns. (Lancaster University )
Listen6:30

Read Story Transcript

Forensic anthropologist Sue Black wants pictures of your hands. 

The Lancaster University professor is leading a project to develop a computer algorithm that police can use to identify child pornographers.

"Very rarely when someone is perpetrating a crime, do they film themselves or photograph themselves doing it. If you go to rob a bank, you don't film yourself robbing a bank. But in terms of child abuse, often the perpetrator will include a part of their anatomy in that image," Black told As It Happens host Carol Off.

"You never see faces for obvious reasons. But hands, people forget about it."

Machine learning 

And there's a lot of information that can be gleaned from hands, Black said. 

Every set is unique, and not only because of the fingerprints.

There are distinct folds and creases, vein patterns, freckles and other skin markings that can be used to identify an individual.

Black said her team already works with authorities on about 50 cases a year to help identify alleged child pornographers using images of their hands in police evidence. 

But each case means hours and hours spent analyzing very disturbing imagery. 

"I'm sure you can imagine the work that we do is very laborious. It takes a long time to be able to look at these images, to abstract all of the anatomical features that we see, whether vein patterns or skin patterns or skin creases," she said.

That's why they want to train computers to do the heavy lifting by using machine learning to develop an algorithm that can memorize a whole array of identifying traits and scan hundreds of images at once. 

"To be able to train a machine to do that, we need to have some really good quality images which we will have to train it," she said.

That's where regular, everyday citizens come in.

Images will be anonymous and not for police use 

Once the team is fully up and running in six to eight weeks, they'll be asking volunteers to send them photographs of their hands — just regular images from mobile phones.

Black estimates they'll need about 5,000 to start. 

The researchers will give each submitted image a two-week cooling off period before they upload it to the database, Black said, giving folks an opportunity to withdraw consent if they change their minds.

Once it goes into the database, Black said the images will be stripped of all identifying data.

"Because we don't really care who they belong to. We're just looking for the extent of variation that you will find in the human hands," she said.

Sue Black is a forensic anthropologist from Lancaster University who is heading up a new project called H-unique to develop a machine learning algorithm to identify child predators based on images of their hands. ( Lancaster University )

Police departments and other authorities won't have access to the images either, she said.

They're strictly for research purposes — to help the scientists build and test an algorithm, which authorities can later use to scan their own databases of images.

"And when you have a large database, we physically cannot look at that number of photographs individually. We need to have computer assistance to do it," she said. 

 'They will be able to link perpetrators across the globe'

"Interpol have millions of images of indecent abuse against children. And so you can imagine they might pick out the perpetrator's hands on a set of images that they found on a computer somewhere in Malaysia. And then six months later, they find the same perpetrator from a set of images that are coming out of Germany," Black said.

"For the first time for the police, they will be able to link perpetrators across the globe. So if people move, we can still track them."

The European Research Council has provided €2.5m ($3.7 Cdn) to fund the five-year project, called H-unqiue, by Lancaster University in collaboration with the University of Dundee.

While Black admits it's possible some abusers will catch on and take more care to hide their hands, she suspects most will not — and even so, any progress made catching predators, even if temporary, is good.

"It's always about the children. Everything is about the children," she said.

"If you can take somebody out of circulation who is abusing children, then what you're doing is hopefully safeguarding the ones that they haven't got to yet."

Written by Sheena Goodyear. Interview with Sue Black produced by Sarah Cooper.

Comments

To encourage thoughtful and respectful conversations, first and last names will appear with each submission to CBC/Radio-Canada's online communities (except in children and youth-oriented communities). Pseudonyms will no longer be permitted.

By submitting a comment, you accept that CBC has the right to reproduce and publish that comment in whole or in part, in any manner CBC chooses. Please note that CBC does not endorse the opinions expressed in comments. Comments on this story are moderated according to our Submission Guidelines. Comments are welcome while open. We reserve the right to close comments at any time.