Volumetric video is a very interesting new medium for filmmaking. In this video, I show how I’ve been using machine learning (AI) to help create live-action volumetric video from stereoscopic sources.
The cameras I used to shoot the footage in the video are the Kandao Ego (https://ift.tt/FplbBKD), and a custom variable-baseline stereo rig.
The Stereo Magnification paper can be found here: https://ift.tt/t14sEAU
And the open-source code can be found here: https://ift.tt/SoZFlBv#volumetric #filmmaking #hologram #machinelearning #viewsynthesis #ai #stereoscopic #lookingglass #kandao #vr #googleai
via YouTube
Harry's Blog 