ContentWise and Vionlabs partner
October 21, 2021
ContentWise and Vionlabs have agreed a partnership that combines their respective personalization and video analysis solutions.
ContentWise, a pioneer in UX personalization for video operators, helps media brands deliver rich, bespoke, and smart digital customer experiences at every touchpoint. ContentWise’s AI personalization technology is used by OTT and streaming operators to process usage data and content metadata to orchestrate and personalize, in real-time, the UI and user experience of millions of subscribers, while providing a unique, individual experience to each.
Vionlabs utilizes computer vision and machine learning to automatically generate previously unavailable video fingerprinting data such as mood. With its AI video content analysis technology, Vionlabs extracts granular metadata from videos in 16 different dimensions, such as scene mood (e.g. “captivating” or “high-octane”) and binge markers. Vionlabs’ technology enriches content metadata by making the elusive emotional gradient that images and sound convey explicit and measurable.
Combined, the two technologies enable unprecedented levels of recommendation accuracy and new, distinctive personalization use cases. In particular:
- Vionlabs mood categories, coupled with ContentWise dynamic categories, which automatically create personalized micro-genre collections for every user, power mood-based personalized recommendations. Users can enjoy very specific and individual genres based on their moods and past behaviors;
- Thanks to ContentWise’s advanced management of serialized and episodic content, and Vionlabs’ automatic AI-detection of intro, recap and credit, operators are able to readily deploy smooth, seamless binge-watching use cases;
- They overcome the lack of rich metadata by automatically generating meta-information processing video frames for any content and by introducing potent new similarity attributes. As a result, personalized recommendation use cases can better foster serendipitous discovery thanks to surprising and unusual selections a user might not have otherwise discovered.
After running detailed experiments, smartly combining and integrating Vionlabs fingerprints to a standard set of metadata, ContentWise finds an increment in accuracy improvement up to 83% in the typical scenario seen at most operators. Plus, the solutions synergies contribute up to 56% catalogue coverage improvements.