Advanced Discovery patented its facial recognition technology, originally created for a law firm client and now available on e-discovery workflows.
If you weren’t already intrigued by the ability of companies like Facebook and Google to recognize the faces of friends and connections in photographs, imagine that capability on a much larger scale. Advanced Discovery is betting it has figured out a way to do this, deploying an algorithm that can recognize tens of thousands of faces. What’s more, the company has already made the technology available through popular e-discovery platforms like kCura’s Relativity and Ipro’s Eclipse.
Advanced Discovery is confident in the novelty of its technology. As part of a larger strategy around patenting what it defined in a statement as “an entirely new class” of technology, Advanced Discovery announced on July 5 that it has filed a patent for its facial recognition software.
Facial technology in legal isn’t entirely new. The Jury Lab, for example, employs facial recognition software to analyze the emotions of potential jurors during jury selection, while such software’s use by the social media company Snapchat led to a lawsuit. Advanced Discovery, however, has created something that can go through thousands upon thousands of faces in images spanning a period of years, Rick Hutchinson, chief technology officer at Advanced Discovery, told Legaltech News.
Hutchinson said the idea for the technology came from a law firm, who was representing a corporate client that was trying to go through about 10 TB of video information to disprove the plaintiff’s claim that they were at a series of corporate events that occurred over a span of years. This video footage captured the faces of “tens of thousands of people,” Hutchinson said, and the defense “had all these images and no way to go through them in a cost-effective manner.”
“If you load 10 TB of images and have 100 reviewers sit down, that can be very costly to have people look,” he added.
The law firm asked Advanced Discovery to handle this task, and from there, Hutchinson “met with the development team, whiteboarded out a solution, and we beta tested it and ran multiple iterations.” This led him to think, “There has to be other cases and situations like this. [So] we built a flexible software platform where it can integrate into industry-known platforms such as Relativity.”
Advanced Discovery isn’t the first discovery solutions provider to dabble in facial recognition technology. Nuix announced in May 2016 that its Nuix 7 processing engine was using facial recognition technology for filtering images in an investigation, while LTU Technologies’ facial recognition software has been deployed for e-discovery purposes.
Hutchinson, however, said that what makes Advanced Discovery unique is that it’s built to integrate with SQL back end review platforms, which comprise the more popular e-discovery workflow tools.
“People are probably doing this ad hoc or figuring out a way to do it, and it’s not really integrated, and on the other side there are software platforms that do this specifically but not for service providers,” he said. “They’re missing the boat.”
The patent filed is part of a broader strategy around filing patents around newer e-discovery technology. Advanced Discovery said it had also filed patents around an analytics platform (which was granted a patent), a data normalization tool, and a technology that automated the elimination of e-discovery data discrepancies.
As to whether the patent will make it harder for other companies to innovate, Hutchinson noted it may complicate efforts for providers wanting to build into integrated e-discovery workflows. “There are multiple ways to do this, yes, but I think they’re going to have to recreate the wheel.”