[CVPR 2024 Highlight] Official repository for the paper "3DGStream: On-the-fly Training of 3D Gaussians for Efficient Streaming of Photo-Realistic Free-Viewpoint Videos".
You need to write a custom script and extract the frames from a multi-view video dataset like N3DV dataset or MeetRoom dataset as we described in the paper.
where i can get this scene dataset in your readmemd which listed as follows