Today we are announcing a major technology update to Microsoft PhotoDNA, the image-matching technology used by organizations and agencies around the world to help fight the spread of child sexual abuse material (aka child pornography) online. Today’s update significantly improves the speed of the PhotoDNA hashing and matching, and expands support to new platforms and programming languages. With its first formal update, PhotoDNA customers will see hashing speeds improve by up to 4 times and matching speeds by 10-20 times and, for the first time, PhotoDNA technology will be provided for Linux and OS X platforms using C#, C++ and Python programming languages.
This week marks the sixth anniversary of PhotoDNA since its launch in December 2009. Since its introduction, PhotoDNA has been made available free of charge to qualified companies, organizations, and developers. Today, more than 100 companies worldwide, including Facebook, Twitter, Kik and Flipboard, use PhotoDNA to automatically detect millions of known illegal child sexual exploitation photos and quickly stop their redistribution online. Law enforcement agencies around the world also use PhotoDNA to aid in their child exploitation investigations, including groundbreaking efforts such as Project Vic.
Today’s announcement builds on other recent investments in PhotoDNA. In March, we announced that PhotoDNA would be available in Azure with our new PhotoDNA Cloud Service, enabling organizations to harness the power of the cloud for a more simple, secure and scalable means of deploying PhotoDNA.
Of course, although today is a positive step forward, there remains much more work to be done to address the problem of online child exploitation. For example, Microsoft remains committed to working with others across the technology industry to drive the kind of scalable, adaptable, and secure advancements needed in video hashing that we’ve been able to achieve together with still images.
None of our PhotoDNA work could happen without partnership from industry leaders and advocates. We especially rely on partnerships with the National Center for Missing and Exploited Children (NCMEC) and the International Center for Missing and Exploited Children (ICMEC) to ensure that PhotoDNA evolves to meet new and changing needs in the fight against exploitation material. Richard Brown, law enforcement technology liaison for ICMEC, describes our collaboration on this update as a “game changer in the effort to expand distribution of PhotoDNA across the globe and support law enforcement’s fight against child sexual exploitation.”
Current PhotoDNA partners will receive information regarding these updates in early 2016. To learn more about PhotoDNA, visit the Microsoft Digital Crimes Unit Newsroom.
PhotoDNA uses “hash” matching technology to automatically detect known illegal photos even if someone has altered them. It converts images into a grayscale format, creating a grid and assigning a numerical value to each tiny square. Those numerical values represent the “hash” of an image, or its “PhotoDNA signature.” The program doesn’t look at images or scan photos; it simply matches a numerical hash against a database of known illegal images, allowing companies to compare millions of photos against a hash set of known illegal child sexual abuse images. The hash set is created by NCMEC and derived from the “worst of the worst” child pornography images uploaded to the CyberTipline by electronic service providers. With PhotoDNA, the industry can work together to more proactively detect these illegal images and disrupt their distribution.