Anttwo / SuGaR

[CVPR 2024] Official PyTorch implementation of SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering

Home Page:https://anttwo.github.io/sugar/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Question about dataset

JiatengLiu opened this issue · comments

commented

Hello! this is a wonderful work!!!
I am a realllly newcomer in Computer Graphics so I have some stupid and a little ridiculous questions about the repo:

  1. I have ran the script and finally get the obj file, through visualization, I find that the result seems to have a lot of messy gauss outside the center of the object, is this normal?
  2. Does the images need to be used only in Gaussian Splatting and not in regularization training and subsequent processes?

the config in my experiment is GeForce RTX 3090 and I test the repo with dataset tandt_db/tandt/truck. Run codes is below:
python gaussian_splatting/train.py -s tandt_db/tandt/truck --iterations 7000 -m output/truck
python train.py -s tandt_db/tandt/truck -c output/truck/ -r "density" --refinement_time "short"
the result is below:
image