search

autonomousvision / monosdf

[NeurIPS'22] MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction
MIT License
573 stars 53 forks source link

The depth and normal about TnT #25

Closed lzhnb closed 2 years ago

lzhnb commented 2 years ago

Thanks for your amazing work!

I have a few questions about the depths and normals for the TnT dataset.

The origin size of images in TnT is [1080, 1920], and the size of your provided images and estimated depth&normal maps is [1152, 2048], why do you scale the image?

Otherwise, I run the script preprocess/extract_monocular_cues.py to estimate depth&normal map, but it only works on the [384, 384] images due to the limitation of the omnidata.

So, how can you obtain the depth&normal maps for the TnT dataset?

Good luck!

lzhnb commented 2 years ago

Ok, I found this in your supplementary: image

niujinshuchong commented 2 years ago

@lzhnb We do have a subsection for this in the paper but not the part you highlighted. Please check C.3 in our updated arxiv paper.

lzhnb commented 2 years ago

@lzhnb We do have a subsection for this in the paper but the part you highlighted. Please check C.3 in our updated arxiv paper.

Thanks for your quick reply!