Single-View 3D Reconstruction via SO(2)-Equivariant Gaussian Sculpting Networks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929496870879232 |
|---|---|
| author | Xu, Ruihan Opipari, Anthony Mah, Joshua Lewis, Stanley Zhang, Haoran Guo, Hanzhe Jenkins, Odest Chadwicke |
| author_facet | Xu, Ruihan Opipari, Anthony Mah, Joshua Lewis, Stanley Zhang, Haoran Guo, Hanzhe Jenkins, Odest Chadwicke |
| contents | This paper introduces SO(2)-Equivariant Gaussian Sculpting Networks (GSNs) as an approach for SO(2)-Equivariant 3D object reconstruction from single-view image observations.
GSNs take a single observation as input to generate a Gaussian splat representation describing the observed object's geometry and texture. By using a shared feature extractor before decoding Gaussian colors, covariances, positions, and opacities, GSNs achieve extremely high throughput (>150FPS). Experiments demonstrate that GSNs can be trained efficiently using a multi-view rendering loss and are competitive, in quality, with expensive diffusion-based reconstruction algorithms. The GSN model is validated on multiple benchmark experiments. Moreover, we demonstrate the potential for GSNs to be used within a robotic manipulation pipeline for object-centric grasping. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_07245 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Single-View 3D Reconstruction via SO(2)-Equivariant Gaussian Sculpting Networks Xu, Ruihan Opipari, Anthony Mah, Joshua Lewis, Stanley Zhang, Haoran Guo, Hanzhe Jenkins, Odest Chadwicke Computer Vision and Pattern Recognition Robotics This paper introduces SO(2)-Equivariant Gaussian Sculpting Networks (GSNs) as an approach for SO(2)-Equivariant 3D object reconstruction from single-view image observations. GSNs take a single observation as input to generate a Gaussian splat representation describing the observed object's geometry and texture. By using a shared feature extractor before decoding Gaussian colors, covariances, positions, and opacities, GSNs achieve extremely high throughput (>150FPS). Experiments demonstrate that GSNs can be trained efficiently using a multi-view rendering loss and are competitive, in quality, with expensive diffusion-based reconstruction algorithms. The GSN model is validated on multiple benchmark experiments. Moreover, we demonstrate the potential for GSNs to be used within a robotic manipulation pipeline for object-centric grasping. |
| title | Single-View 3D Reconstruction via SO(2)-Equivariant Gaussian Sculpting Networks |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2409.07245 |