RCDINO: Enhancing Radar–Camera 3D Object Detection with DINOv2 Semantic Features
Optical Memory and Neural Networks · 2025-12 · 34 (S1) · S47-S57
DOI: 10.3103/s1060992x25601708 ↗Kurz gefasst
Fachartikel in der Zeitschrift Optical Memory and Neural Networks, online erschienen am 19.12.2025. Verfasst von O. Matykina und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 31 Literaturangaben.
Abstract (Originalsprache)
Three-dimensional object detection is essential for autonomous driving and robotics, relying on effective fusion of multimodal data from cameras and radar. This work proposes RCDINO, a multimodal transformer-based model that enhances visual backbone features by fusing them with semantically rich representations from the pretrained DINOv2 foundation model. This approach enriches visual representations and improves the model’s detection performance while preserving compatibility with the baseline architecture. Experiments on the nuScenes dataset demonstrate that RCDINO achieves state-of-the-art performance among radar–camera models, with 56.4 NDS and 48.1 mAP. Our implementation is available at https://github.com/OlgaMatykina/RCDINO.
Zitieren
APA
Matykina, O., & Yudin, D. (2025). RCDINO: Enhancing Radar–Camera 3D Object Detection with DINOv2 Semantic Features. Optical Memory and Neural Networks, 34(S1), S47-S57. https://doi.org/10.3103/s1060992x25601708
Vancouver
Matykina O, Yudin D. RCDINO: Enhancing Radar–Camera 3D Object Detection with DINOv2 Semantic Features. Optical Memory and Neural Networks. 2025;34(S1):S47-S57. doi:10.3103/s1060992x25601708
BibTeX
@article{Matykina2025,
author = {Matykina, O. and Yudin, D.},
title = {RCDINO: Enhancing Radar–Camera 3D Object Detection with DINOv2 Semantic Features},
journal = {Optical Memory and Neural Networks},
year = {2025},
volume = {34},
number = {S1},
pages = {S47--S57},
doi = {10.3103/s1060992x25601708}
}
EUVIMED