arXiv (Cornell University)· 2020· Preprint
Multiview Neural Surface Reconstruction by Disentangling Geometry and\n Appearance
- 345citations
- 2020year
Short summary
A novel neural network simultaneously reconstructs 3D object geometry and appearance from multiple 2D images, achieving state-of-the-art fidelity.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces a neural network architecture for multiview 3D surface reconstruction.
- Simultaneously learns geometry, camera parameters, and a neural renderer.
- Geometry represented as a zero level-set of a neural network.
- Neural renderer models various lighting and material conditions.
- Achieves state-of-the-art results on the DTU MVS dataset.
AI-generated from the title and abstract; the full text is not read.
Abstract
In this work we address the challenging problem of multiview 3D surface\nreconstruction. We introduce a neural network architecture that simultaneously\nlearns the unknown geometry, camera parameters, and a neural renderer that\napproximates the light reflected from the surface towards the camera. The\ngeometry is represented as a zero level-set of a neural network, while the\nneural renderer, derived from the rendering equation, is capable of\n(implicitly) modeling a wide set of lighting conditions and materials. We\ntrained our network on real world 2D images of objects with different material\nproperties, lighting conditions, and noisy camera initializations from the DTU\nMVS dataset. We found our model to produce state of the art 3D surface\nreconstructions with high fidelity, resolution and detail.\n
The authors' abstract, as published at the source. arXiv (Cornell University), 2020 · 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