About the depth loss and volume rendering sampling strategy

[anhquancao]

Hello, Thank you for your interesting work! I found your formulation of depth loss with volume rendering really cool. I was wondering if you have tried combining both of them, or combining depth loss with semantic loss? Also, what sampling strategy did you use for the volume rendering?

[GANWANSHUI]

Hi, thank you very much for your attention and insightful question.

We are trying to combine the depth loss and the classification loss and will update it if we have some useful findings. We think the rendering fashion should also be suitable for training with semantic loss. We may investigate it recently but welcome to try it.

The sampling strategy is similar to the regular nerf in voxel format, like this one: https://github.com/sunset1995/DirectVoxGO

[anhquancao]

Hello, thank you for your information and the reference!