search

kwea123 / nerf_pl

NeRF (Neural Radiance Fields) and NeRF in the Wild using pytorch-lightning
https://www.youtube.com/playlist?list=PLDV2CyUo4q-K02pNEyDr7DYpTQuka3mbV
MIT License
2.74k stars 483 forks source link

Beta in the NeRF-w #108

Open Panxuran opened 3 years ago

Panxuran commented 3 years ago

Hi,

Thank you for contributing this repo.

My understanding on the beta in the NeRF-w is computed independently for each location, and also rendered with similar alpha-compositing weights. I found similar code in your repo in the models/rendering.py file line 176 results['beta'] = reduce(transient_weights*transient_betas, 'n1 n2 -> n1', 'sum')

However, since the paper claims beta as the variance of each location, should the corresponding scaling coefficient be the square of the alpha-compositing weights? results['beta'] = reduce(transient_weights*transient_weights*transient_betas, 'n1 n2 -> n1', 'sum')

I might made a mistake, and I really look forward to your reply.

Thanks in advance for all the help you can provide.