Using AI for Up-Conversion
Originally Aired - Monday, April 17 | 11:20 AM - 12:00 PM PT
This Session Has Not Started Yet
Be sure to come back after the session starts to have access to session resources.
There are many existing libraries of valuable content, from movies through to sports content, which are not available in the resolutions that today’s consumer has readily available. Traditional approaches to up-converting this content would have either involved a vey expensive manual process, or relatively simplistic spatio-temporal filtering, resulting in more pixels, but not more detail. AI has the potential to fundamentally change the possibilities for up-conversion, by creating plausible content rather than just filtering.
In this talk, we describe the type of architecture and results of such an AI-based approach, providing a new option for increasing the potential value of existing assets.