Analytics Are Changing the Fight Against Child Exploitation
Author: Louis F. Quijas

Senior Customer Advisor, Cellebrite
Former FBI Assistant Director & NA Session #168

Technology is evolving faster than ever before. Today, if you want to communicate with family and friends, share pictures with loved ones, browse on the internet, or upload a post on social media, it is second nature to just whip out a mobile device and do it all from the palm of your hand.

Unfortunately, for all their benefits, always-on, connected devices and networks have created channels for predators to exploit children and proliferate explicit material. Nearly every week, the media reports on yet another tragic victim of child sexual exploitation (CSE) and abuse. The number of photos and videos seized and reviewed annually is staggering. In a 2016 report by the US Department of Justice, The National Center for Mission & Exploited Children (NCMEC) estimated that more than 26 million sexual abuse images and videos were reviewed by their analysts in 2015 alone. That number continues to climb exponentially each year, challenging law enforcement agencies around the globe.

 “When I started in forensics, the majority of devices we seized were desktops and laptops,” said Det. Randy Kyburz, Certified Digital Forensic Examiner with the Seattle Police Internet Crimes Against Children Unit. “Years ago, we’d walk out of a crime scene with maybe one of each. Today, we often collect 30+ devices at a crime scene, with smart phones making up about 40 percent of total devices recovered.”

 Traditional digital forensic workflows combined with sentencing guidelines and the sheer volume of offenders has created an epidemic where child victims are often undetected and undiscovered, and the crimes committed against them are never investigated. The failure of this detection enables the continued access to, and abuse of, these children.

An Urgent Call to Action

Thanks to ubiquitous connectivity, offenders have virtually unlimited access to unsuspecting children and lurid content. An FBI investigation of a single child pornography website hosted on Tor, the anonymous internet network, had approximately 200,000 registered users with 100,000 individuals accessing the site during a 12-day period¹. Individual offenders often possess massive collections of terabytes or even petabytes of data on multiple devices.

 Until recently, when tens of thousands of images of child abuse material were seized by law enforcement, many of those photos or videos were destined to be left on devices, in the cloud and in evidence lockers. However, the tide has begun to turn, as innovative digital forensic tools now provide reliable ways to extract, parse and identify images and video of known and unknown victims. Companies focused on identification, extraction and analysis of digital intelligence continue to invest significantly to unlock, access and analyze this data quickly and

in a forensically sound manner. New analytics algorithms provide powerful options for correlating and analyzing files from various computer, social media, cloud, mobile, cell tower and other digital sources. Full integration with Project VIC, the Child Abuse Image Database (CAID) and other defined hash value databases significantly reduces manual analysis efforts, not to mention the psychological stress of reviewing sensitive material.

Optimizing shared resources and workflows

The goal is steadfast: identify and save more exploited children – quickly - by putting the power back in the hands of those dedicated to protecting children around the world. Video analytics empowers forensics practitioners, investigators and analysts to efficiently manage the growing volume of evidential data and reduce case cycle times.

Today, all case stakeholders can access forensic artifacts and collaborate in real time using the latest tools on the market. Unique machine learning algorithms accelerate time to evidence. The power of video analytics lies not only in the ability to correlate and review actionable insights across all data sources, but also to help quickly find evidence when investigators don’t know what they are looking for – what people are talking about, languages they are using, locations they’ve frequented, etc. Advances in video and image analytics in solutions available today deliver both critical triage capabilities at the scene and more in-depth investigational analysis in the lab.

Specifically, these solutions provide the following benefits:

Accelerate time to evidence with advanced machine learning

Once suspected CSE-related material is obtained through a forensic process, video analytics and CSE image categorization can automatically identify images and videos using machine learning neural network-based algorithms. Once categorized, images can be filtered based on categories such as face, nudity, and suspected child exploitation, so relevant images and videos can be identified quickly.

Quickly identify and cross-match victims with facial detection

Powerful algorithms can now automatically detect faces within any picture or video available to the system, enabling investigators to immediately and accurately cross-match individual faces. This allows investigators to quickly identify additional images or videos of the same victim.

Analyze conversations for potential luring or abuse

Natural language processing goes beyond regex and simple watch lists to uncover names, addresses, locations and more from artifacts like emails, websites, text messages or even images that contain text, using OCR, in multiple languages.

Leverage public domain cloud data to correlate evidence

Visualize and analyze publicly available data from supported social media and cloud-based sources in a unified format to track behavior, uncover common connections and correlate critical evidence that can help build a stronger case.

Seamless integration with Project VIC, CAID and other hash databases

Existence of known incriminating images can be automatically identified by matching image hash values, and can then be classified using pre-defined CSE severity categories. Previously unknown images that are discovered can also be categorized, tagged and exported back to Project VIC and CAID databases in a seamless and integrated process.

A collective, collaborative fight to serve and protect

Preventing child exploitation takes collaboration, real-time information and an ongoing commitment to identify every victim quickly and get criminals – and the content they produce and share – off the streets. With more and more children using mobile devices at an earlier age, the risks are only getting bigger.

Digital intelligence helps law enforcement protect the innocent

Digital data – especially images and video - plays an increasingly important role in investigations and operations of all kinds. Enabling access, sharing and analysis of this digital data from mobile devices, social media, cloud, computer and other sources helps investigators build the strongest cases quickly, even in the most complex situations.

The goal for law enforcement is to find relevant, actionable digital evidence quickly.  Partnering with companies such as Cellebrite for solutions that automate analysis of huge volumes of digital data will help achieve a shared goal: to find and protect exploited children, and make a safer world more possible every day.


1 -

About the Author:

Louis F. Quijas
Louis F. Quijas is a former law enforcement professional who has serve
d at both the federal and local levels.  His storied career includes appointments by the FBI Director to oversee the Office of Law Enforcement, and by the President of the United States, as the Assistant Secretary for the Office for State and Local Law Enforcement at the Department of Homeland Security.  Lou has also served on several national boards - most notably, as President of the National Latino Peace Officers Association, and a member of the Executive Committee of the International Association of Chiefs of Police.  He currently serves on the National Sheriffs Association’s Global Policing Committee.