Amazon today announced new features for its Rekognition video segment detection API that uses machine learning to allow VOD providers to automatically detect frame accurate end credits, black frame segments, shot changes and color bars. The update now includes support for four new types of segments: opening credits, content segments, slates and studio logos. It also improves accuracy for credits and shot detection and new filters to control black frame detection.
Introduced a year ago, Rekognition video segment detection gives VOD providers the ability to automate media analysis tasks, specifically the detection of black frames, end credits, shot changes and color bars, using fully managed, ML-powered APIs. Amazon says Rekognition’s video segment detection capability can help reduce review time from hours to minutes, cut costs of manual operations by up to 5x and prepare archived content for VOD on short turnaround.
For the new features, Rekognition automatically detects frames for opening and closing credits for movies or TV shows and identify the start and end of “slates,” which are sections that appear at the beginning of videos, and include text metadata about the episode, studio, video format, audio channels. This makes it easier for operators to use the metadata for analysis or remove slates when prepping for streaming services.
The update also allows operators to now be able to identify and review production studio logos that appear at the beginning of on-demand content as well as identify the start and end of content segments to review for runtime and customize content in different languages for international markets.
More information is available at Amazon’s blog.
