New IBM Cloud service uses Watson to generate video metadata

The service is designed to help media companies and advertisers better target their video content.


IBM on Monday announced a new Watson-powered cloud service it plans to roll out to analyze video metadata. The service, announced at the National Association of Broadcasters Show, should help content providers and advertisers spot ways to appeal to targeted audiences.

The service will use Watson's cognitive capabilities to extract metadata like keywords, concepts, visual imagery, tone and emotional context. It uses a range of Watson APIs, including: Tone Analyzer, Personality Insights, Natural Language Understanding, and Visual Recognition. It can also segment videos into logical scenes.

For instance, a sports network could use the service to quickly and automatically identify and package happy scenes from basketball related content to be used in advertising -- rather than having a person manually choose video clips to use.

IBM said its range of AI-powered capabilities is unique in the marketplace of video analysis. Given the growing amount of content delivered over the cloud through various devices, it's harder for companies to extract the all of the insights available, the company argues.

"We are seeing that the dramatic growth in multi-screen content and viewing options is creating a critical need for M&E companies to transform the way content is developed and delivered to address evolving audience behaviors," said Steve Canepa, general manager for IBM Global Telecommunications, Media and Entertainment industry, in a statement.

The service, which should be available later this year, builds on IBM's previous research into ways Watson can serve the media and entertainment industry. Last year, IBM used Watson to produce a movie trailer, while earlier this year Watson technology produced a highlight reel of the Master's golf tournament.

IBM puts deep learning to work on diabetic eye disease

IBM puts deep learning to work on diabetic eye disease

More on IBM: