Probabilistic Camera Distortion Correction Using Deep Gaussian Processes
Journal of Imaging · 02.07.2026 · 12 (7) · 296
DOI: 10.3390/jimaging12070296 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Imaging, online erschienen am 02.07.2026. Verfasst von Ivan De Boi und 7 weiteren Autor:innen. Beteiligte Einrichtung: InViLab, University of Antwerp, Groenenborgerlaan 179, 2020 Antwerp, Belgium. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research Foundation - Flanders, FWO postdoc fellowship.
Abstract (Originalsprache)
Accurate lens distortion correction is important for calibration, registration, image stitching, and 3D reconstruction, especially in low-data device-specific settings where disposable or specialised cameras cannot provide large calibration datasets. We address distortion correction for cameras with highly irregular or non-stationary distortion fields, where fixed polynomial models and generic learning-based rectification methods can struggle. We propose a framework based on Deep Gaussian Processes (DGPs) to model the non-linear mapping required for undistortion. The key motivation is that conventional single-layer GPs with stationary kernels must use one global notion of smoothness, whereas DGPs can represent spatially varying behaviour through composed latent mappings while preserving per-pixel predictive uncertainty. This uncertainty can be used to identify or downweight unreliable corrected regions in downstream tasks. We evaluate the method on three real camera datasets with increasing distortion complexity. The full structured acquisitions contain 512 horizontal and 512 vertical line images per camera. These are not thousands of natural calibration images, but they yield up to 29,532, 11,311, and 31,686 detected intersection correspondences for the RPI, Theta, and Pillcam datasets, respectively. This distinction is important for cameras where acquiring many independent images is impractical. The results are assessed using qualitative rectification, uncertainty maps, normalised collinearity errors, and total training time. Polynomial calibration remains strongest for the regular radial RPI distortion, while DGP and DGP2 models show lower normalised collinearity-error distributions than the standard GP and lightweight MLP baselines on the more distorted Theta and Pillcam datasets. For the full datasets, total DGP/DGP2 training times ranged from 2383.50 s to 10092.50 s, reflecting the additional computational cost of probabilistic non-stationary modelling.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- InViLab, University of Antwerp, Groenenborgerlaan 179, 2020 Antwerp, Belgium.
- AI Lab, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium.
- Institute of Systems and Robotics, Department of Electrical and Computer Engineering (DEEC), University of Coimbra, Rua Sílvio Lima, 3030-290 Coimbra, Portugal.
ORCID-Profile: Ivan De Boi, Rhys G Evans, Stuti Pathak, Thomas De Kerf, Marnix Van Soom, Sam Van der Jeught, Helder Araújo, Rudi Penne
Förderung
- Research Foundation - Flanders
- FWO postdoc fellowship
Zitieren
APA
De Boi, I., Evans, R. G., Pathak, S., Kerf, T., Soom, M. V., Jeught, S. V., Araújo, H., & Penne, R. (2026). Probabilistic Camera Distortion Correction Using Deep Gaussian Processes. Journal of Imaging, 12(7), 296. https://doi.org/10.3390/jimaging12070296
Vancouver
De Boi I, Evans RG, Pathak S, Kerf T, Soom MV, Jeught SV, et al. Probabilistic Camera Distortion Correction Using Deep Gaussian Processes. Journal of Imaging. 2026;12(7):296. doi:10.3390/jimaging12070296
BibTeX
@article{DeBoi2026,
author = {De Boi, Ivan and Evans, Rhys G and Pathak, Stuti and Kerf, Thomas De and Soom, Marnix Van and Jeught, Sam Van der and Araújo, Helder and Penne, Rudi},
title = {Probabilistic Camera Distortion Correction Using Deep Gaussian Processes},
journal = {Journal of Imaging},
year = {2026},
volume = {12},
number = {7},
pages = {296},
doi = {10.3390/jimaging12070296}
}
EUVIMED