fdarmon/NeuralWarp

Quantitative results on DTU dataset

Closed this issue · 2 comments

As descripted in your paper, "the results for each method are taken from their original paper".
But why results for Neus[34] are different from original paper?

[34] Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, and Wenping Wang. NeuS: Learning neural implicit surfaces by volume rendering for multi-view reconstruction. In Adv. Neural Inform. Process. Syst., 2021.

Hi, sorry for the late reply.

I reported the results from the version that was available at that time (https://arxiv.org/pdf/2106.10689v1.pdf). I will update the tables with the newer results.

Thanks for your reply