Did you try directly optimize the octree, instead of after initialization with NeRF (or shNeRF) prediction? about plenoctree

[rayleizhu]

By the way, in the readme, you mentioned that:

For example, for the drums scene in the NeRF-Synthetic dataset, 100 / 10000 sampling view directions takes about 2 mins / 2 hours to finish the plenoctree extraction. It produce raw plenoctrees with PSNR=22.49 / 23.84 (before optimization). Note that extraction from a NeRF-SH model produce a raw plenoctree with PSNR=25.01.

How about after optimization? In another word, is the initialization important?

[yangjiheng]

By the way, in the readme, you mentioned that:

For example, for the drums scene in the NeRF-Synthetic dataset, 100 / 10000 sampling view directions takes about 2 mins / 2 hours to finish the plenoctree extraction. It produce raw plenoctrees with PSNR=22.49 / 23.84 (before optimization). Note that extraction from a NeRF-SH model produce a raw plenoctree with PSNR=25.01.

How about after optimization? In another word, is the initialization important?

Well, seems like plenoxels you were talking about.

[rayleizhu]

By the way, in the readme, you mentioned that:

For example, for the drums scene in the NeRF-Synthetic dataset, 100 / 10000 sampling view directions takes about 2 mins / 2 hours to finish the plenoctree extraction. It produce raw plenoctrees with PSNR=22.49 / 23.84 (before optimization). Note that extraction from a NeRF-SH model produce a raw plenoctree with PSNR=25.01.

How about after optimization? In another word, is the initialization important?

Well, seems like plenoxels you were talking about.

I know. It was Sep. 1st, 2021 when I raised the question.

After I asked it here, I did some experiments to find the answer. I did figure out that the initialization is not important, and immediately realized that Alex Yu may already know this. Soon, Plenvoxels is published on Arxiv.