International database to help child abuse investigators

A cloud-based service that will enable law enforcement in the UK to quickly compare seized child abuse content against an international database is to launch in the next six months. The ultimate goal is to help authorities quickly identify the most recent material, to help save victims depicted in that content.

According to the body coordinating the effort, the International Centre for Missing and Exploited Children, the renewed focus on tracking victims rather than abusers has led to 80 children being rescued in the past 18 months.

The system is part of Project Vic, an international collaboration that began with Homeland Security and federal law enforcement divisions in the US, and now includes Interpol and forces in the UK, New Zealand, Canada, Australia and more. Data analytics software from NetClean and Hubstream has been integral to the project’s growth, along with Microsoft’s PhotoDNA.

According to Johann Hofmann, product manager for NetClean Analyse, the project was first conceived nearly two years ago. NetClean had been helping the Swedish police assign “digital fingerprints” to child abuse content for five years when representatives sat down with Microsoft and law enforcement officials in the US. “We presented the work we had been doing and sat down to try to identify some of the challenges law enforcement has,” he told

It quickly became transparent that all the different levels of law enforcement in the US were not sharing information efficiently enough. The caseload of individual officers was becoming too great as the problem increased. It meant every time a suspect was identified and the data seized, it was being processed, known images identified and a sentence for possession administered.

“They then just leave that material aside and move on to the next case,” Hofmann told

The point of Project Vic is to not only synch up information other agencies might have on that same child abuse content, but to help eliminate which pieces of content are old and which might contain new information, so officers can hone in on what’s pertinent, faster.

“Sharing information had not been done in the past,” said Hofmann. “But by sharing they can actually reduce the amount of time they have to spend looking at the same images, to focus on new images and new material — new unidentified victims. The project will aim to ultimately rescue more victims.”

A global outlook for identifying victims and perpetrators is key, as demonstrated by the recent arrest made by Homeland Security in the US of 14 men operating a child abuse website using Tor which had 27,000 members. That investigation identified 251 victims in 39 US states and five other countries, including the UK, Canada, New Zealand, Australia and Belgium. 

The new system is implemented from the moment an investigation begins, when any computers, smartphones or other devices are seized. NetClean Analyse is used to process the data, and any piece of illegal content downloaded or saved is assigned a digital signature that is automatically calculated. These are then uploaded to the database, along with categories and additional information — for instance, if a victim or perpetrator in the image, or their whereabouts, has already been identified. All that data is stored in the cloud — when someone else connects and downloads the data, they can run new images against existing content.

Hofmann explains: “They can bounce fingerprints against this cloud service, and immediately know what’s been seen in the pasts. All different law enforcement agencies across the US and other countries are connected and can push up new fingerprints on a daily basis. And it’s growing fast.

“They can benefit from other work being done all over the country and elsewhere. They also get additional information, like whether the victim or vendor been identified, if the picture comes from a known image series that’s been/being investigated, potentially internationally.”

The cloud aspect of the system is still in the testing phase, and has been for around a year — NetClean and other parties involved (Homeland Security coordinates it with the International Centre for Missing and Exploited Children, with help from others in the industry) want it to be “100 percent” before the roll out, which is due to happen in the next six months.

According to Richard Brown, Law Enforcement Liaison at the International Centre for Missing and Exploited Children, the system has already helped police “pre-identify” around 85 percent of images seized on a hard drive. “Instead the investigator can concentrate on the remaining 15 percent of images never seen by law enforcement,” Brown tells “Project Vic wants to make sure these images are examined and not just sent to the evidence vault. Each agency in the project has vowed to make a difference and ‘Leave No Child Behind’. This is our motto and we feel with new technologies combined with our new approach we can save a child’s life.” 

“It will have a huge impact because of the amount of people involved,” says Hofmann of the roll out. “We’re trying to get other vendors to adopt the protocol and to share information.”

Brown explains that other countries that would like to join Project Vic can do so easily, because it was “built on technology that is very open to connectivity”. He adds: “In the last 18 months more than 80 children have been rescued because of the new focus on the victim centric approach to investigation.”

Microsoft’s Photo DNA and binary hashing techniques are used to match content. However, NetClean is currently looking into a new, robust technology that would be the video equivalent of Photo DNA. The details are currently under wraps, but Hofmann did say: “It will have a massive impact — video is an incredibly huge problem because of the amount of information you have there. It’s very hard for law enforcement to look for all that information: it’s tough on the eyes and takes time. But if you can find a robust system, it will save a lot of work and also mental health.”

What began as a small consortium of interested parties has now expanded to the point where Hofmann says “we can’t even have all the logos on one slide anymore”. Part of Project Vic involves training law enforcement to use the tech, and training those individuals to become trainers themselves — thus the network is rapidly flourishing.

According to Hofmann, it couldn’t have come soon enough.

“It is an increasing problem. The amount of content being produced and shared is growing exponentially. More and more people have smartphones and digital cameras, it’s so easy to share online on a closed network and storing it is really cheap — you can buy terabytes of disk space for no money. Criminals help each other online with how to commit crimes and get around police officers and their methods. So they are definitely becoming more sophisticated. That’s why it’s so important for law enforcement and industry to work together to develop novel technologies.”

The UK government has been piling the pressure on search engines, and in particular Google, to share the burden of tackling the growing problem of child abuse content online. Google already worked with the Internet Watch Foundation to takedown and report identified content, and last year executive chairman Eric Schmidt pledged to do more, including implementing an algorithm tweak that would limit search results for suspected child sexual abuse queries using Microsoft’s PhotoDNA. 

Of course, it’s entirely possible those sharing the content would quickly become savvy and use alternate language and references to find material. Nevertheless Hofmann says industry’s role is a vital and growing one. “I do think we’re seeing change at this point in the UK,” he said. “It has to start with law enforcement, which then typically puts more pressure on industry to do something. You can see the change all over the world. Facebook, for instance, is doing tonnes of work in this area.

“It’s such a huge problem, and everyone thinks it’s an important problem we need to combat. We just have to make companies aware.”