EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Feature-Level Vehicle-Infrastructure Cooperative Perception with Adaptive Fusion for 3D Object Detection

Shuangzhi Yu, Jiankun Peng, Shaojie Wang, Di Wu, Chunye Ma

Smart Cities · 14.10.2025 · 8 (5) · 171

DOI: 10.3390/smartcities8050171 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Smart Cities, online erschienen am 14.10.2025. Verfasst von Shuangzhi Yu und 4 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: the Key R&D Program of Shandong Province, China, the SEU Innovation Capability Enhancement Plan for Doctoral Students of southeast university.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As vehicle-centric perception struggles with occlusion and dense traffic, vehicle-infrastructure cooperative perception (VICP) offers a viable route to extend sensing coverage and robustness. This study proposes a feature-level VICP framework that fuses vehicle- and roadside-derived visual features via V2X communication. The model integrates four components: regional feature reconstruction (RFR) for transferring region-specific roadside cues, context-driven channel attention (CDCA) for channel recalibration, uncertainty-weighted fusion (UWF) for confidence-guided weighting, and point sampling voxel fusion (PSVF) for efficient alignment. Evaluated on the DAIR-V2X-C benchmark, our method consistently outperforms state-of-the-art feature-level fusion baselines, achieving improved AP3D and APBEV (reported settings: 16.31% and 21.49%, respectively). Ablations show RFR provides the largest single-module gain +3.27% AP3D and +3.85% APBEV, UWF yields substantial robustness gains, and CDCA offers modest calibration benefits. The framework enhances occlusion handling and cross-view detection while reducing dependence on explicit camera calibration, supporting more generalizable cooperative perception.

Förderung

  • the Key R&D Program of Shandong Province, China
  • the SEU Innovation Capability Enhancement Plan for Doctoral Students of southeast university

Zitieren

APA

Yu, S., Peng, J., Wang, S., Wu, D., & Ma, C. (2025). Feature-Level Vehicle-Infrastructure Cooperative Perception with Adaptive Fusion for 3D Object Detection. Smart Cities, 8(5), 171. https://doi.org/10.3390/smartcities8050171

Vancouver

Yu S, Peng J, Wang S, Wu D, Ma C. Feature-Level Vehicle-Infrastructure Cooperative Perception with Adaptive Fusion for 3D Object Detection. Smart Cities. 2025;8(5):171. doi:10.3390/smartcities8050171

BibTeX
@article{Yu2025,
  author = {Yu, Shuangzhi and Peng, Jiankun and Wang, Shaojie and Wu, Di and Ma, Chunye},
  title = {Feature-Level Vehicle-Infrastructure Cooperative Perception with Adaptive Fusion for 3D Object Detection},
  journal = {Smart Cities},
  year = {2025},
  volume = {8},
  number = {5},
  pages = {171},
  doi = {10.3390/smartcities8050171}
}
Kostenlos starten →