search

hustvl / 4DGaussians

[CVPR 2024] 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering
https://guanjunwu.github.io/4dgs/
Apache License 2.0
2.25k stars 187 forks source link

How to output dense point cloud? #130

Open cwwjyh opened 7 months ago

cwwjyh commented 7 months ago

Thanks for the great work!

How can I get dynamic point cloud? image in the above code, store the static point cloud on different training iterations. It is very sparse, as below. image

So I want to know how to get dense point cloud when the dynamic Gaussian is played after training. This point cloud will store output/my_checkpoint_dir

Looking forward to your reply!

guanjunwu commented 6 months ago

Hi, do you mean to get each timestamps' 3D Gaussians after training? export per frame 3DGS as written in readme.md may help you. such as: python export_perframe_3DGS.py --iteration 14000 --configs arguments/dnerf/bouncingballs.py --model_path output/dnerf/bouncingballs