encountering error at coarse mesh extraction

[AbnerVictor]

Hi, I encountered the following error when I was trying to run the second stage of your code, here is the error message.

Could you kindly give me some suggestions to fix this problem? Thanks!

To be clear, I use the Nerf synthetic dataset and generate the gs checkpoint using your code.

Traceback (most recent call last): File "train.py", line 136, in coarse_mesh_args = AttrDict({ File "SuGaR-main/sugar_extractors/coarse_mesh.py", line 410, in extract_mesh_from_coarse_sugar o3d_bg_mesh, o3d_bg_densities = o3d.geometry.TriangleMesh.create_from_point_cloud_poisson( RuntimeError: [Open3D Error] (static std::tuple<std::shared_ptropen3d::geometry::TriangleMesh, std::vector<double, std::allocator > > open3d::geometry::TriangleMesh::CreateFromPointCloudPoisson(const open3d::geometry::PointCloud&, size_t, float, float, bool, int)) /root/Open3D/cpp/open3d/geometry/SurfaceReconstructionPoisson.cpp:732: Point cloud has no normals

[Anttwo]

Hello AbnerVictor,

Thank you for noticing this!

This may be due to an empty background, are you using a synthetic dataset? As I explain in this issue (https://github.com/Anttwo/SuGaR/issues/19), since we were interested in full, real scenes only, there was always a background in my scenes and I forgot to handle the case of an empty background, that could be useful for synthetic datasets. I'm sorry for that!

I just changed the coarse_mesh.py script. I did not re-test the code because this change is really minor (it was just a few 'if's to add), but it should work with empty backgrounds now. Feel free to tell me if you meet another error related to this.

Thank you!

[AbnerVictor]

Hello AbnerVictor,

Thank you for noticing this!

This may be due to an empty background, are you using a synthetic dataset? As I explain in this issue (#19), since we were interested in full, real scenes only, there was always a background in my scenes and I forgot to handle the case of an empty background, that could be useful for synthetic datasets. I'm sorry for that!

I just changed the coarse_mesh.py script. I did not re-test the code because this change is really minor (it was just a few 'if's to add), but it should work with empty backgrounds now. Feel free to tell me if you meet another error related to this.

Thank you!

Thank you. I suppose it is because of the empty background, I'm using a synthetic dataset with a pure black background. I will try your new code!

It is a great work!

[yuedajiong]

My case is without background, like this:

[Freeverc]

It occured on drone collected dataset by 231218 commit :c2d1c1c6839dd96fcd153a1bd129a138e5dee79e Loading the coarse SuGaR model from path ./output/coarse/1121222212333331/sugarcoarse_3Dgs7000_sdfestim02_sdfnorm02/15000.pt... Use min to initialize scales. Initialized radiuses for 3D Gauss Rasterizer Coarse model loaded. Coarse model parameters: _points torch.Size([183672, 3]) True all_densities torch.Size([183672, 1]) True _scales torch.Size([183672, 3]) True _quaternions torch.Size([183672, 4]) True _sh_coordinates_dc torch.Size([183672, 1, 3]) True _sh_coordinates_rest torch.Size([183672, 15, 3]) True Number of gaussians: 183672 Opacities min/max/mean: tensor(5.1140e-14, device='cuda:0') tensor(1., device='cuda:0') tensor(0.9421, device='cuda:0') Quantile 0.0: 5.114017323222357e-14 Quantile 0.1: 0.7956185936927795 Quantile 0.2: 0.9892562031745911 Quantile 0.3: 0.9988507032394409 Quantile 0.4: 0.9995555281639099 Quantile 0.5: 0.999764621257782 Quantile 0.6: 0.9998601675033569 Quantile 0.7: 0.9999138116836548 Quantile 0.8: 0.9999490976333618 Quantile 0.9: 0.9999756813049316

Starting pruning low opacity gaussians... WARNING! During optimization, you should use a densifier to prune low opacity points. This function does not preserve the state of an optimizer, and sets requires_grad=False to all parameters. Number of gaussians left: 176499 Opacities min/max/mean: tensor(0.5001, device='cuda:0') tensor(1., device='cuda:0') tensor(0.9697, device='cuda:0') Quantile 0.0: 0.5001183748245239 Quantile 0.1: 0.896555483341217 Quantile 0.2: 0.996113657951355 Quantile 0.3: 0.9991563558578491 Quantile 0.4: 0.9996246099472046 Quantile 0.5: 0.999788224697113 Quantile 0.6: 0.9998703002929688 Quantile 0.7: 0.9999186992645264 Quantile 0.8: 0.9999513626098633 Quantile 0.9: 0.999976634979248 Processing frame 0/84... Current point cloud for level 1.0 has 0 points. Processing frame 30/84... Current point cloud for level 1.0 has 0 points. Processing frame 60/84... Current point cloud for level 1.0 has 0 points.

========== Processing surface level 1.0 ========== Final point cloud for level 1.0 has 0 points. Using default, camera based bounding box. Centering bounding box. Foreground points: torch.Size([0, 3]) torch.Size([0, 3]) torch.Size([0, 3]) Background points: torch.Size([0, 3]) torch.Size([0, 3]) torch.Size([0, 3])

-----Foreground mesh----- Computing points, colors and normals... Cleaning Point Cloud... Finished computing points, colors and normals. Now computing mesh... Traceback (most recent call last): File "/home/code/SuGaR/train.py", line 145, in coarse_mesh_path = extract_mesh_from_coarse_sugar(coarse_mesh_args)[0] File "/home/code/SuGaR/sugar_extractors/coarse_mesh.py", line 387, in extract_mesh_from_coarse_sugar o3d_fg_mesh, o3d_fg_densities = o3d.geometry.TriangleMesh.create_from_point_cloud_poisson( RuntimeError: [Open3D Error] (static std::tuple<std::shared_ptropen3d::geometry::TriangleMesh, std::vector<double, std::allocator > > open3d::geometry::TriangleMesh::CreateFromPointCloudPoisson(const open3d::geometry::PointCloud&, size_t, float, float, bool, int)) /root/Open3D/cpp/open3d/geometry/SurfaceReconstructionPoisson.cpp:732: Point cloud has no normals

[hughkhu]

same problem. How to solve it?