About renderer

[ydwxuexi55]

I use the renderer according to the usage, but always get a black background image, what is the reason?

[twowwj]

I met the same problem. Anybody can help?

[a4152684]

same issue

[a4152684]

我怀疑是11.x的cuda无法正常运行p2i

[boqiuphd]

I use the renderer according to the usage, but always get a black background image, what is the reason?

Need to make sure the point cloud data is loaded correctly onto the cuda, if the point cloud data is on the CPU it will return an empty projection map. data = torch.from_numpy(point_cloud).unsqueeze(0).cuda()

[MMMaverick]

I use the renderer according to the usage, but always get a black background image, what is the reason?

Need to make sure the point cloud data is loaded correctly onto the cuda, if the point cloud data is on the CPU it will return an empty projection map. data = torch.from_numpy(point_cloud).unsqueeze(0).cuda()

That work for me, thank you!

[MMMaverick]

@boqiuphd But I found that the visualization of point clouds for various categories looks quite strange, as if they are disordered. Do you have any suggestions?

[boqiuphd]

@boqiuphd But I found that the visualization of point clouds for various categories looks quite strange, as if they are disordered. Do you have any suggestions?

Are the projection maps for all viewpoints weird? Or maybe you can modify class ComputeDepthMaps in SpareNet/utils/p2i_utils.py, for example, to customize a function to generate the eyes_pos_list you need.

[MMMaverick]

It was probably my incorrect use of reshape that caused the data structure to get messed up. Thank you for your help！

[boqiuphd]

It was probably my incorrect use of reshape that caused the data structure to get messed up. Thank you for your help！

You're welcome and good luck with your research!