diff-gaussian-rasterization icon indicating copy to clipboard operation
diff-gaussian-rasterization copied to clipboard

Results 14 diff-gaussian-rasterization issues
Sort by recently updated
recently updated
newest added

Thank you for the amazing work. I encounter a problem regarding the shape of the rendered depth map. I tried to retrieve the rendered_image and rendered_depth by running: `rendered_image, radii,...

How to render multi-channel feature image using gaussian rasterization? For example, colors_precomp is a 8-dim vector represented feature of point cloud, how to render the feature map(8, h, w) without...

Thanks for this wonderful repo! I wonder if you can support normal map rendering. Since normal map can show more details.

1. does this make trained scene representation/rendering better than original 3DGS rasterizer? 2. is there a way to use this to make loss term used in 3DGS more efficient and...

The official version of 3D Gaussian Splatting only uses Color Loss + SSIM Loss during their training. But now there's your re-written diff-gaussian-rasterization. I would like to ask, is it...

Hi, thank you so much for your great repository! I have a question about depth estimation. I have installed your modified version of the differentiable rasterizer and noticed that depth...

Hi I got a problem while I was installing the error is shown below (DreamG) E:\2.code\dreamgaussian>pip install ./diff-gaussian-rasterization Processing e:\2.code\dreamgaussian\diff-gaussian-rasterization Preparing metadata (setup.py) ... done Building wheels for collected packages:...

Hi @ashawkey, thanks for your beautiful repositories (Stable DreamFusion, DreamGaussian, etc)! I'm heavily using your DreamGaussian repository and would like to know if you could implement gradients for the camera...

In forward.cu line 250, you calculate the depth as: ``` depths[idx] = p_view.z; ``` And p_view is obtained in line 190 ``` in_frustum(idx, orig_points, viewmatrix, projmatrix, prefiltered, p_view) ``` where...

感谢您精彩的工作! 有个困惑是,您实现了对于depth和α的渲染,以及相应的backward cuda实现,但是如果在外层python中,如果我只做了depth的监督,而没有做α的监督,这样在loss.backward的时候会报错吗(或者会导致没被监督的α计算错误吗)? 再次感谢您精彩的工作,期待您的回复!