Wid3R: Wide Field-of-View 3D Reconstruction via Camera Model Conditioning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914425742557184 |
|---|---|
| author | Jung, Dongki Choi, Jaehoon Qureshi, Adil Jeong, Somi Manocha, Dinesh Yeon, Suyong |
| author_facet | Jung, Dongki Choi, Jaehoon Qureshi, Adil Jeong, Somi Manocha, Dinesh Yeon, Suyong |
| contents | We present Wid3R, a feed-forward neural network for multi-view visual geometry reconstruction that supports wide field-of-view camera models. Unlike existing methods that assume rectified or pinhole inputs, Wid3R directly models wide-angle imagery without explicit calibration or undistortion. Our approach leverages a ray-based representation with spherical harmonics and introduces a novel camera model token to enable distortion-aware reconstruction. To the best of our knowledge, Wid3R is the first multi-frame feed-forward 3D reconstruction method that supports 360 imagery. Moreover, we show that conditioning on diverse camera types improves generalization to 360 scenes and alleviates data sparsity issues. Wid3R achieves significant performance gains, improving AUC@30 by up to +33.67 on Zip-NeRF (fisheye) and +77.33 on Stanford2D3D (360). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_05321 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Wid3R: Wide Field-of-View 3D Reconstruction via Camera Model Conditioning Jung, Dongki Choi, Jaehoon Qureshi, Adil Jeong, Somi Manocha, Dinesh Yeon, Suyong Computer Vision and Pattern Recognition We present Wid3R, a feed-forward neural network for multi-view visual geometry reconstruction that supports wide field-of-view camera models. Unlike existing methods that assume rectified or pinhole inputs, Wid3R directly models wide-angle imagery without explicit calibration or undistortion. Our approach leverages a ray-based representation with spherical harmonics and introduces a novel camera model token to enable distortion-aware reconstruction. To the best of our knowledge, Wid3R is the first multi-frame feed-forward 3D reconstruction method that supports 360 imagery. Moreover, we show that conditioning on diverse camera types improves generalization to 360 scenes and alleviates data sparsity issues. Wid3R achieves significant performance gains, improving AUC@30 by up to +33.67 on Zip-NeRF (fisheye) and +77.33 on Stanford2D3D (360). |
| title | Wid3R: Wide Field-of-View 3D Reconstruction via Camera Model Conditioning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.05321 |