Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

show result in UE #26

Open
luoww1992 opened this issue May 24, 2024 · 5 comments
Open

show result in UE #26

luoww1992 opened this issue May 24, 2024 · 5 comments

Comments

@luoww1992
Copy link

luoww1992 commented May 24, 2024

i want to show it in UE,but i notice it need original .ply file
how to translate the result to original .ply file ?
or some others advices。

@jianglh-WHU
Copy link
Contributor

Sorry, in the present, the final results from octree-gs cannot be directly converted to the 3D-GS format due to the implementation of the MLP.

@luoww1992
Copy link
Author

@jianglh-WHU
if without mlp, it can do ?
if yes, how to do ?

@ZhiyeTang
Copy link

@luoww1992 Maybe you can try pruning all view-dependent information used in Octree-GS and re-train the code. Afterward, you can use the way as in generate_neural_gaussians() to generate all neural Gaussians into a .ply file as what it looks like in vanilla 3DGS. Theoretically, it is compatible with the existing UE plugins for 3DGS.

@ZhiyeTang
Copy link

ZhiyeTang commented Jul 21, 2024

The fundamental reason why Scaffold-GS derivatives (like Octree-GS and HAC) are difficult to import into UE is because they require generating neural Gaussians on-the-fly every time the view changes. Therefore, with pruning such view-dependent information, such anchor-based methods can be correctly converted to "static" 3DGS format, and loaded into UE like real vanilla 3DGS.

@jianglh-WHU
Copy link
Contributor

Yeah! That's a solid move, switching Octree-GS to 3DGS. Just gotta keep a few things in mind:

  1. Shift the opacity mask and swap the activation function from tanh to sigmoid when predicting the opacity of the neural Gaussian.
  2. It's better to predict spherical harmonics instead of color.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants