Generative Data Augmentation for ArUco-Free RGB-Based 6-DoF Object Pose Estimation
Journal of Imaging · 29.05.2026 · 12 (6) · 244
DOI: 10.3390/jimaging12060244 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Imaging, online erschienen am 29.05.2026. Verfasst von Carmelo Scribano und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Physics, Informatics and Mathematics, University of Modena and Reggio Emilia, Via Campi 213/B, 41125 Modena, Italy. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Gruppo Nazionale per il Calcolo Scientifico, European Union-FSE-REACT-EU, PON Research and Innovation.
Abstract (Originalsprache)
In recent years, data-driven approaches have become increasingly important in industrial computer vision applications, particularly for 6-Degrees-of-Freedom (6-DoF) object pose estimation. However, benchmark datasets may unintentionally introduce biases that affect the reliability of learned models. In this work, we investigate the shortcut bias induced by fiducial ArUco markers in the widely used Linemod dataset. Although such markers are typically absent in real industrial environments, they introduce unintended visual cues that neural networks tend to exploit. As a result, model selection based on state-of-the-art benchmarks can be biased, since the reported performance often reflects reliance on these shortcuts rather than on robust feature extraction. Using saliency map analysis, we show that often a large portion of the model’s attention is concentrated on these markers, revealing the presence of a shortcut that artificially boosts pose estimation performance. To mitigate this issue, we propose a data augmentation pipeline based on generative AI techniques that removes the markers and replaces the background with more realistic synthesized scenes. Experimental results indicate a noticeable drop in performance when models trained on the original Linemod dataset are evaluated in ArUco-free environments, confirming the presence of background-induced biases. Training with the proposed generative-swapped dataset leads to improved robustness and better generalization to unseen scenarios, although it does not fully eliminate the problem. Overall, the results highlight the impact of background-related biases in pose estimation benchmarks and suggest that the proposed augmentation strategy represents a practical and scalable step toward developing more reliable 6-DoF pose estimation systems for industrial applications, while leaving room for further improvements.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Physics, Informatics and Mathematics, University of Modena and Reggio Emilia, Via Campi 213/B, 41125 Modena, Italy.
- HIPERT s.r.l, Via Inventori, 37, 41121 Modena, Italy.
ORCID-Profile: Carmelo Scribano, Giorgia Franchini, Elena Govi, Davide Sapienza, Tobia Poppi, Micaela Verucchi, Marko Bertogna
Förderung
- Gruppo Nazionale per il Calcolo Scientifico
- European Union-FSE-REACT-EU
- PON Research and Innovation
Zitieren
APA
Scribano, C., Ferrari, I., Franchini, G., Govi, E., Sapienza, D., Poppi, T., Verucchi, M., & Bertogna, M. (2026). Generative Data Augmentation for ArUco-Free RGB-Based 6-DoF Object Pose Estimation. Journal of Imaging, 12(6), 244. https://doi.org/10.3390/jimaging12060244
Vancouver
Scribano C, Ferrari I, Franchini G, Govi E, Sapienza D, Poppi T, et al. Generative Data Augmentation for ArUco-Free RGB-Based 6-DoF Object Pose Estimation. Journal of Imaging. 2026;12(6):244. doi:10.3390/jimaging12060244
BibTeX
@article{Scribano2026,
author = {Scribano, Carmelo and Ferrari, Iacopo and Franchini, Giorgia and Govi, Elena and Sapienza, Davide and Poppi, Tobia and Verucchi, Micaela and Bertogna, Marko},
title = {Generative Data Augmentation for ArUco-Free RGB-Based 6-DoF Object Pose Estimation},
journal = {Journal of Imaging},
year = {2026},
volume = {12},
number = {6},
pages = {244},
doi = {10.3390/jimaging12060244}
}
EUVIMED