These Magic Moments: Differentiable Uncertainty Quantification of Radiance Field Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912285423828992 |
|---|---|
| author | Ewen, Parker Chen, Hao Isaacson, Seth Wilson, Joey Skinner, Katherine A. Vasudevan, Ram |
| author_facet | Ewen, Parker Chen, Hao Isaacson, Seth Wilson, Joey Skinner, Katherine A. Vasudevan, Ram |
| contents | This paper introduces a novel approach to uncertainty quantification for radiance fields by leveraging higher-order moments of the rendering equation. Uncertainty quantification is crucial for downstream tasks including view planning and scene understanding, where safety and robustness are paramount. However, the high dimensionality and complexity of radiance fields pose significant challenges for uncertainty quantification, limiting the use of these uncertainty quantification methods in high-speed decision-making. We demonstrate that the probabilistic nature of the rendering process enables efficient and differentiable computation of higher-order moments for radiance field outputs, including color, depth, and semantic predictions. Our method outperforms existing radiance field uncertainty estimation techniques while offering a more direct, computationally efficient, and differentiable formulation without the need for post-processing. Beyond uncertainty quantification, we also illustrate the utility of our approach in downstream applications such as next-best-view (NBV) selection and active ray sampling for neural radiance field training. Extensive experiments on synthetic and real-world scenes confirm the efficacy of our approach, which achieves state-of-the-art performance while maintaining simplicity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_14665 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | These Magic Moments: Differentiable Uncertainty Quantification of Radiance Field Models Ewen, Parker Chen, Hao Isaacson, Seth Wilson, Joey Skinner, Katherine A. Vasudevan, Ram Computer Vision and Pattern Recognition Robotics This paper introduces a novel approach to uncertainty quantification for radiance fields by leveraging higher-order moments of the rendering equation. Uncertainty quantification is crucial for downstream tasks including view planning and scene understanding, where safety and robustness are paramount. However, the high dimensionality and complexity of radiance fields pose significant challenges for uncertainty quantification, limiting the use of these uncertainty quantification methods in high-speed decision-making. We demonstrate that the probabilistic nature of the rendering process enables efficient and differentiable computation of higher-order moments for radiance field outputs, including color, depth, and semantic predictions. Our method outperforms existing radiance field uncertainty estimation techniques while offering a more direct, computationally efficient, and differentiable formulation without the need for post-processing. Beyond uncertainty quantification, we also illustrate the utility of our approach in downstream applications such as next-best-view (NBV) selection and active ray sampling for neural radiance field training. Extensive experiments on synthetic and real-world scenes confirm the efficacy of our approach, which achieves state-of-the-art performance while maintaining simplicity. |
| title | These Magic Moments: Differentiable Uncertainty Quantification of Radiance Field Models |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2503.14665 |