Reuters experiments with face recognition technology to search and index video
The technology, devised by Viewdle, analyses each frame of video footage, looking 'inside the clip' to identify people
The technology, devised by Viewdle, analyses each frame of video footage, looking 'inside the clip' to identify people
This article was migrated from an old version of our website in 2025. As a result, it might have some low-quality images or non-functioning links - if there's any issues you'd like to see fixed, get in touch with us at info@journalism.co.uk.
There used to be something here that couldn't be migrated - please contact us at info@journalism.co.uk if you'd like to see this updated!
Reuters is experimenting with facial-recognition technology that could allow it to index and make fully searchable its video news content.
The technology, devised by Viewdle, analyses each frame of video footage, looking 'inside the clip' to identify the appearance of people on-screen.
The new technology deal comes just a day after Reuters announced that it had forged a deal with Attributor to use its content monitoring technology to track the distribution of online news across social media sites.
Developers say the facial-recognition technology is valuable in powering internal indexing for content providers and enhancing audience-facing search and advertising capabilities.
Enter a search name and with each return the technology automatically cues the video to the first section where your chosen person appears, lists how long they appear for in that segment and where else they appear in the clip and for how long.
The search also returns information on other people recognised in the clips.
The technology appears to work best in close up shots. A search for ' Tom Cruise ' returns a single clip with three designated segments in which he appears. He actually appears a fourth time, albeit in a longer shot where he is not as easily definable.
The search return for this clip also tells the user that film director Cameron Crowe appears in the video, although the technology fails to list either Kanye West and Orlando Bloom - people recognised in other content - both of who also appear fleetingly in the clip.
However, the technology is largely impressive and Viewdle says it's building the world's biggest people-in-video reference database and has multiple patents in preparation.
The Reuters Labs, which showcases emerging technologies and tools being tested by the news giant, is hosting the technology.
"Our goal at Reuters Labs is to pro-actively identify emerging technologies that add value to our trusted news offerings and keep Reuters at the forefront of the evolution of digital media," said Nic Fulton, chief scientist of Reuters.
"By implementing facial recognition capabilities throughout our consumer-facing video offerings with Viewdle, we can study the benefits of this search technology with an eye to using it with other Reuters video products."