Isaaclabe / DGD-Dynamic-3D-Gaussians-Distillation

Official implementation of "DGD: Dynamic 3D Gaussians Distillation".
https://isaaclabe.github.io/DGD-Website/
MIT License
48 stars 1 forks source link

How to render in novel views #4

Open Drow999 opened 6 days ago

Drow999 commented 6 days ago

Thank you for sharing this great project! I'm trying to render the model trained with Dinov2, Could you clarify how parameters like novel views, points and theta should be determined when rendering?

Isaaclabe commented 3 days ago

Hi,

As it is explain in the Readme, the novel view, points and theta are determined in the command line prompt:

python render.py -s path/to/your/dataset -m output/exp-name --fundation_model "DINOv2" --semantic_dimension 384 --iterations 40_000 --frame k --novel_views i --points "(x1,y1)" "(x2,y2)" --thetas "ϴ1" "ϴ2"