2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)· 2022
Point-NeRF: Point-based Neural Radiance Fields
- 526citations
- 2022year
Short summary
Point-NeRF models radiance fields using neural 3D point clouds, achieving NeRF-level visual quality with 30x faster training and enabling efficient rendering by aggregating features near scene surfaces.
AI-generated from the title and abstract; the full text is not read.
Key points
- Models radiance fields using neural 3D point clouds with associated neural features.
- Achieves NeRF-level visual quality with 30x faster training time.
- Enables efficient rendering via aggregation of neural point features near scene surfaces.
- Can be initialized by direct inference from a pre-trained network.
- Includes a pruning and growing mechanism to handle errors from other 3D reconstruction methods.
AI-generated from the title and abstract; the full text is not read.
Abstract
Volumetric neural rendering methods like NeRF [34] generate high-quality view synthesis results but are optimized per-scene leading to prohibitive reconstruction time. On the other hand, deep multi-view stereo methods can quickly reconstruct scene geometry via direct network inference. Point-NeRF combines the advantages of these two approaches by using neural 3D point clouds, with associated neural features, to model a radiance field. Point-NeRF can be rendered efficiently by aggregating neural point features near scene surfaces, in a ray marching-based rendering pipeline. Moreover, Point-NeRF can be initialized via direct inference of a pre-trained deep network to produce a neural point cloud; this point cloud can be finetuned to surpass the visual quality of NeRF with 30× faster training time. Point-NeRF can be combined with other 3D re-construction methods and handles the errors and outliers in such methods via a novel pruning and growing mechanism.
The authors' abstract, as published at the source. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Computer Graphics and Computer-Aided Design
Computer Graphics and Computer-Aided DesignComputer Science