NVlabs / neuralangelo

Official implementation of "Neuralangelo: High-Fidelity Neural Surface Reconstruction" (CVPR 2023)
https://research.nvidia.com/labs/dir/neuralangelo/
Other
4.38k stars 388 forks source link

Performance Variations when Reproducing Results on DTU Dataset #65

Open Dangzheng opened 1 year ago

Dangzheng commented 1 year ago

Hello and thank you for the insightful work you've shared!

While attempting to reproduce your results on the DTU dataset, I noticed some variations in performance.

Below is a result summarization:

image

Interestingly, only case 106 aligns with the paper's results, while the others seem to diverge.

Implementation Details:

  1. I've used the provided Docker file, with a minor modification: pulling image FROM nvcr.io/nvidia/pytorch:22.12-py3 due to my NVIDIA driver version (11.8).

  2. The evaluation method was in line with the NeuS paper. The alignment of case 106's results with the paper gives me confidence in the accuracy of my evaluation.

Questions:

  1. Might there be specific package versions that are critical to achieving the results? If so, could you kindly share those details?

  2. Are there other aspects or considerations that might explain the variations in performance?

Your guidance would be immensely appreciated. Thank you in advance for your time and expertise!

mli0603 commented 1 year ago

Hi @Dangzheng @Runsong123

Thanks for reporting. We will look into this. A few quick comments:

jk4011 commented 10 months ago

Hi @mli0603 , Any update for mesh release?