Temporal Preservation for Volumetric Content
Alexandria Porter
Unknown
- 0 Collaborators
This tool is a solution for viewing dynamic and high quality volumetric content without the overhead of heavy infrastructure increasing user interaction. ...learn more
Project status: Published/In Market
Virtual Reality, Mobile, Internet of Things, Game Development, Artificial Intelligence, Graphics and Media
Intel Technologies
Intel Python,
Intel GPA
Overview / Usage
Volumetric video is an extremely expensive process because every single frame is its own asset. Aside from mass batch optimization there are few solutions out there that can allow users to keep the same detail all while providing a small disc size footprint. Utilization of Temporal Preservation allows this content to be in a smaller footprint all while increasing visual fidelity, surface adherence, and textural quality. While the cost of typical strategies like mass batch optimization have a linear increase in disc foot print where each frame costs a set amount Temporal Preservation has an increase up front while asset increase over time is minimal. This is much more akin to typical asset utilization in both games and movies, where an objects cost in mesh and texture is high while animation assets have little increase in the overall footprint. In games footprint tend to be more conscientious of packet size since individual users need to download assets. Movies and render farms in general are becoming more space conscious since the increased use of SSDs becoming prevalent and the average disc size has decreased while thru-put has increased. Using smaller file sizes allows for increased processing speed especially where discs are centralized like a data center. This also allows for fully 3 dimensional streaming content where render to texture programs would typically be used light weight meshes can be sent across allowing users to interact within VR, AR, or MR without heavy data center structure on the delivering end.
Methodology / Approach
Video capture > mesh generation > edge cleanup > optimization
Utilizing the same topology overtime.
Technologies Used
Photogrammetry, Intel Xenon processors, Intel Python, Visual Studio, GPA, NUC, Up Board, VR, AR, MR, Unity, Maya, Photoshop,