search

zju3dv / manhattan_sdf

Code for "Neural 3D Scene Reconstruction with the Manhattan-world Assumption" CVPR 2022 Oral
https://zju3dv.github.io/manhattan_sdf/
Other
500 stars 35 forks source link

Quantify the results #60

Closed 05063112lcs closed 5 months ago

05063112lcs commented 5 months ago

Hello, thank you very much for your excellent work, I would like to know where to get some quantitative data indicators after running your project?

ghy0324 commented 5 months ago

Hi! Thanks for your interest! When we conducted experiments on ScanNet, since there was sensor depth data in the dataset, we were able to obtain ground truth geometry through RGB-D fusion, which allowed us to perform quantitative evaluation of the reconstruction results from our RGB-based method. For data without ground truth depth, it may not be possible to quantitatively evaluate the geometric reconstruction results.