TensoRF
TensoRF copied to clipboard
apchenstu
[ECCV 2022] Tensorial Radiance Fields, a novel approach to model and reconstruct radiance fields
Hey I created a custom dataset with blender from the original blender files, I face to following issue: ``` Iteration 04000: train_psnr = 17.97 test_psnr = 0.00 mse = 0.016522:...
Hi, In [issue-46](https://github.com/apchenstu/TensoRF/issues/46#issuecomment-1369576671), you mention that tensorf uses a different coordinate system. Can you please elaborate on that? From the code, I understand that NeRF uses the standard `(x,y,z)` system,...
Can you explain what does parameter near_far means? Or maybe provide where i can get some informations about this, since I couldn't find any explanation in both your code and...
Thank you for your great work! When I try to use TensoRF in indoor scene dataset, such as room 0 in [pre-rendered Replica data](https://www.dropbox.com/sh/9yu1elddll00sdl/AAC-rSJdLX0C6HhKXGKMOIija?dl=0) used in [Semantic-NeRF](https://github.com/Harry-Zhi/semantic_nerf), it seems not...
Hello Chen, thanks for your excellent work! I'm writing to ask some question about the combination of the line vector and the plane mat in density pipeline. Why this process...
When creating the AlphaGridMask [here](https://github.com/apchenstu/TensoRF/blob/main/models/tensorBase.py#L333), the alpha grid that is passed is of shape `(size_z, size_y, size_x)` instead of `(size_x, size_y, size_z)` because of the transpose operation [here](https://github.com/apchenstu/TensoRF/blob/main/models/tensorBase.py#L325). However, when...
When computing the depth map via volume rendering, the following line is executed at https://github.com/apchenstu/TensoRF/blob/main/models/tensorBase.py#L463 ``` depth_map = depth_map + (1. - acc_map) * rays_chunk[..., -1] ``` Why is this...
Why is it that after I use the Mish activation function instead of the Relu function in the color decoder network, no color information is extracted for data such as...
Hello, Thanks for your great job sharing with us, could you please help me with getting access to the model weights during the training? I see in rendering there is...
Hi, I noticed you're computing `nSamples` in two different places with slightly different logic. In [`utils.py`](https://github.com/apchenstu/TensoRF/blob/main/utils.py#L59--L60), it is simple logic. We need two samples for every unit length of voxels,...