Cloud giant's AI software will 'recognise' famous guests at Prince Harry and Meghan Markle's big day
Saturday will see Prince Harry marry Meghan Markle at a celebrity-packed St. George's Chapel in Windsor - and thanks to Amazon Web Services (AWS), no famous face will go unrecognised.
The cloud giant is using the live broadcast of the royal wedding to showcase its facial recognition technology, working with Sky News to provide viewers with a detailed explanation of who each famous person is and how they know the royal couple.
Rekognition will power the Sky News app's 'Who's Who Live' feature that will automatically highlight celebrities as they appear on screen.
Users of the Sky News app will be able to immediately see onscreen captions and graphics for each recognised famous guest and navigate the data without leaving the app, while keeping the livestream of the wedding on their screen.
AWS powers Sky News app's 'Who's Who Live' feature/Credit: AWS Elemental
"We're excited to have helped Sky News successfully build and deploy the Royal Wedding: Who's Who Live app," said ALex Dunlap, general manager of AWS Elemental, which is AWS's division that specialises in creating videos at scale.
"The high visibility and unpredictable audience size for this type of event made AWS cloud services, including those for media, a great solution by giving the ability to test quickly, only pay for what was used, and produce a reliable, high-quality experience in a matter of weeks."
A video feed from an outside broadcast van located near the chapel will capture the faces of the rich and famous as they arrive, feeding the information to an AWS Elemental live small form factor appliance, located nearby, and ingest this real-time data into an entirely cloud-based workflow.
Sky News is also using data analysis platform GrayMeta, which has combined with Amazon Rekognition video and image analysis service, to provide the app with a tag of each celebrity.
It's the ideal opportunity for AWS to showcase its technology - hopefully it performs better than the police's own systems, which managed to incorrectly match faces in crowds with those from a criminal database 95% of the time, according to a privacy organisation's report.