EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Scenario-Driven Synthetic Data Generation Framework for Visual Perception Evaluation Under Adverse Driving Conditions

Wei Xu iD, Dominique Gruyer iD, Alexandra Duminil iD, Sio-Song Ieng iD

Sensors · 30.05.2026 · 26 (11) · 3464

DOI: 10.3390/s26113464 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 30.05.2026. Verfasst von Wei Xu und 3 weiteren Autor:innen. Beteiligte Einrichtung: Perceptions, Interactions, Behaviour and Simulations of Road and Street Users Laboratory (PICS-L), Department of Components and Systems (COSYS), Gust…. Die Arbeit nennt 85 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union’s Horizon 2021 research and innovation programme.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Developing and evaluating visual perception systems for autonomous vehicles requires data across diverse and adverse driving conditions, yet collecting and annotating such real-world data is costly and often impractical. To address this challenge, we propose a modular, scenario-driven framework for generating synthetic datasets tailored to the evaluation of visual perception functions. The framework aligns with the operational boundaries and detection–response requirements of automated driving functions and comprises three stages: (1) configuring use-case-driven scenarios, (2) generating sensor data and ground truth via simulation, and (3) post-processing to ensure dataset usability. Designed to be generic and flexible, the framework is instantiated and demonstrated through its integration with specific platforms and tools, namely Pro-SiVIC and RTMaps. We evaluate the generated dataset from two perspectives, image fidelity and perception performance under synthetic weather conditions, in comparison to real-world conditions. Furthermore, we train multiple perception models under different learning paradigms, including baseline, transfer-learning, and mixed-training strategies, to examine the influence of synthetic data on robustness. Experimental results demonstrate not only the high quality of the generated data but also its effectiveness in evaluating visual perception functions, as well as its benefit to model robustness and generalization.

Schlagwörter

Stichwörter der Autor:innen

ODD and OEDRadverse conditionsautonomous drivingscenario-based frameworksynthetic datasetsvisual perception evaluation

Autor:innen und Institutionen

  • Perceptions, Interactions, Behaviour and Simulations of Road and Street Users Laboratory (PICS-L), Department of Components and Systems (COSYS), Gustave Eiffel University, 77454 Marne-la-Vallée, France.
  • Doctoral School of Information and Communication Sciences and Technologies (ED STIC), Paris-Saclay University, 91190 Gif-sur-Yvette, France.

ORCID-Profile: Wei Xu, Dominique Gruyer, Alexandra Duminil, Sio-Song Ieng

Förderung

  • European Union’s Horizon 2021 research and innovation programme

Zitieren

APA

Xu, W., Gruyer, D., Duminil, A., & Ieng, S. S. (2026). Scenario-Driven Synthetic Data Generation Framework for Visual Perception Evaluation Under Adverse Driving Conditions. Sensors, 26(11), 3464. https://doi.org/10.3390/s26113464

Vancouver

Xu W, Gruyer D, Duminil A, Ieng SS. Scenario-Driven Synthetic Data Generation Framework for Visual Perception Evaluation Under Adverse Driving Conditions. Sensors. 2026;26(11):3464. doi:10.3390/s26113464

BibTeX
@article{Xu2026,
  author = {Xu, Wei and Gruyer, Dominique and Duminil, Alexandra and Ieng, Sio-Song},
  title = {Scenario-Driven Synthetic Data Generation Framework for Visual Perception Evaluation Under Adverse Driving Conditions},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {11},
  pages = {3464},
  doi = {10.3390/s26113464}
}
Kostenlos starten →