FIORD: A Fisheye Indoor-Outdoor Dataset with LIDAR Ground Truth for 3D Scene Reconstruction and Benchmarking
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909572638179328 |
|---|---|
| author | Gunes, Ulas Turkulainen, Matias Ren, Xuqian Solin, Arno Kannala, Juho Rahtu, Esa |
| author_facet | Gunes, Ulas Turkulainen, Matias Ren, Xuqian Solin, Arno Kannala, Juho Rahtu, Esa |
| contents | The development of large-scale 3D scene reconstruction and novel view synthesis methods mostly rely on datasets comprising perspective images with narrow fields of view (FoV). While effective for small-scale scenes, these datasets require large image sets and extensive structure-from-motion (SfM) processing, limiting scalability. To address this, we introduce a fisheye image dataset tailored for scene reconstruction tasks. Using dual 200-degree fisheye lenses, our dataset provides full 360-degree coverage of 5 indoor and 5 outdoor scenes. Each scene has sparse SfM point clouds and precise LIDAR-derived dense point clouds that can be used as geometric ground-truth, enabling robust benchmarking under challenging conditions such as occlusions and reflections. While the baseline experiments focus on vanilla Gaussian Splatting and NeRF based Nerfacto methods, the dataset supports diverse approaches for scene reconstruction, novel view synthesis, and image-based rendering. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_01732 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FIORD: A Fisheye Indoor-Outdoor Dataset with LIDAR Ground Truth for 3D Scene Reconstruction and Benchmarking Gunes, Ulas Turkulainen, Matias Ren, Xuqian Solin, Arno Kannala, Juho Rahtu, Esa Computer Vision and Pattern Recognition The development of large-scale 3D scene reconstruction and novel view synthesis methods mostly rely on datasets comprising perspective images with narrow fields of view (FoV). While effective for small-scale scenes, these datasets require large image sets and extensive structure-from-motion (SfM) processing, limiting scalability. To address this, we introduce a fisheye image dataset tailored for scene reconstruction tasks. Using dual 200-degree fisheye lenses, our dataset provides full 360-degree coverage of 5 indoor and 5 outdoor scenes. Each scene has sparse SfM point clouds and precise LIDAR-derived dense point clouds that can be used as geometric ground-truth, enabling robust benchmarking under challenging conditions such as occlusions and reflections. While the baseline experiments focus on vanilla Gaussian Splatting and NeRF based Nerfacto methods, the dataset supports diverse approaches for scene reconstruction, novel view synthesis, and image-based rendering. |
| title | FIORD: A Fisheye Indoor-Outdoor Dataset with LIDAR Ground Truth for 3D Scene Reconstruction and Benchmarking |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.01732 |