Inria

Framestore's Superman talk details its 4D Gaussian splatting pipeline, from a 192-camera, 88-second take trained in Postshot to per-shot PLY files in Houdini.

aussian Light Transport fits global illumination as a mixture of 13D Gaussians, training 10 to 23 times faster than Neural Radiosity in a few megabytes.

The original implementation is receiving major training speed boosts.

In early September, we took at a look at Improving NeRF Quality by Progressive Camera Placement by Georgios Kopanas and George Drettakis from Inria.