Normal-guided Detail-Preserving Neural Implicit Function for High-Fidelity 3D Surface Reconstruction
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916700267479040 |
|---|---|
| author | Patel, Aarya Laga, Hamid Sharma, Ojaswa |
| author_facet | Patel, Aarya Laga, Hamid Sharma, Ojaswa |
| contents | Neural implicit representations have emerged as a powerful paradigm for 3D reconstruction. However, despite their success, existing methods fail to capture fine geometric details and thin structures, especially in scenarios where only sparse multi-view RGB images of the objects of interest are available. This paper shows that training neural representations with first-order differential properties (surface normals) leads to highly accurate 3D surface reconstruction, even with as few as two RGB images. Using input RGB images, we compute approximate ground-truth surface normals from depth maps produced by an off-the-shelf monocular depth estimator. During training, we directly locate the surface point of the SDF network and supervise its normal with the one estimated from the depth map. Extensive experiments demonstrate that our method achieves state-of-the-art reconstruction accuracy with a minimal number of views, capturing intricate geometric details and thin structures that were previously challenging to capture. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_04861 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Normal-guided Detail-Preserving Neural Implicit Function for High-Fidelity 3D Surface Reconstruction Patel, Aarya Laga, Hamid Sharma, Ojaswa Computer Vision and Pattern Recognition Graphics I.3.5 Neural implicit representations have emerged as a powerful paradigm for 3D reconstruction. However, despite their success, existing methods fail to capture fine geometric details and thin structures, especially in scenarios where only sparse multi-view RGB images of the objects of interest are available. This paper shows that training neural representations with first-order differential properties (surface normals) leads to highly accurate 3D surface reconstruction, even with as few as two RGB images. Using input RGB images, we compute approximate ground-truth surface normals from depth maps produced by an off-the-shelf monocular depth estimator. During training, we directly locate the surface point of the SDF network and supervise its normal with the one estimated from the depth map. Extensive experiments demonstrate that our method achieves state-of-the-art reconstruction accuracy with a minimal number of views, capturing intricate geometric details and thin structures that were previously challenging to capture. |
| title | Normal-guided Detail-Preserving Neural Implicit Function for High-Fidelity 3D Surface Reconstruction |
| topic | Computer Vision and Pattern Recognition Graphics I.3.5 |
| url | https://arxiv.org/abs/2406.04861 |