Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2404.02155 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914758050971648 |
|---|---|
| author | Ahn, Joshua Wang, Haochen Yeh, Raymond A. Shakhnarovich, Greg |
| author_facet | Ahn, Joshua Wang, Haochen Yeh, Raymond A. Shakhnarovich, Greg |
| contents | Scale-ambiguity in 3D scene dimensions leads to magnitude-ambiguity of volumetric densities in neural radiance fields, i.e., the densities double when scene size is halved, and vice versa. We call this property alpha invariance. For NeRFs to better maintain alpha invariance, we recommend 1) parameterizing both distance and volume densities in log space, and 2) a discretization-agnostic initialization strategy to guarantee high ray transmittance. We revisit a few popular radiance field models and find that these systems use various heuristics to deal with issues arising from scene scaling. We test their behaviors and show our recipe to be more robust. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_02155 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Alpha Invariance: On Inverse Scaling Between Distance and Volume Density in Neural Radiance Fields Ahn, Joshua Wang, Haochen Yeh, Raymond A. Shakhnarovich, Greg Computer Vision and Pattern Recognition Scale-ambiguity in 3D scene dimensions leads to magnitude-ambiguity of volumetric densities in neural radiance fields, i.e., the densities double when scene size is halved, and vice versa. We call this property alpha invariance. For NeRFs to better maintain alpha invariance, we recommend 1) parameterizing both distance and volume densities in log space, and 2) a discretization-agnostic initialization strategy to guarantee high ray transmittance. We revisit a few popular radiance field models and find that these systems use various heuristics to deal with issues arising from scene scaling. We test their behaviors and show our recipe to be more robust. |
| title | Alpha Invariance: On Inverse Scaling Between Distance and Volume Density in Neural Radiance Fields |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.02155 |