Enhanced Real‐Time Traffic Sign Recognition via Lightweight Neural Networks and Wavelet Transform
IET Intelligent Transport Systems · 2026-01 · 20 (1)
DOI: 10.1049/itr2.70286 ↗Kurz gefasst
Fachartikel in der Zeitschrift IET Intelligent Transport Systems, online erschienen am 13.07.2026. Verfasst von Yi‐Chao Wu und 2 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Traffic sign recognition plays a pivotal role in enhancing road safety and supporting driver assistance systems. This paper introduces a lightweight neural network model, integrating ShuffleNet V2 with discrete wavelet transform (DWT), designed for real‐time traffic sign recognition. Leveraging ShuffleNet V2's efficiency and DWT's ability to capture fine‐grained visual features, our model achieves high recognition accuracy with minimal computational resources. Extensive experiments under diverse weather and lighting conditions demonstrate the model's robustness and adaptability. Here, we show an improvement in small target detection accuracy by 5.8% compared to baseline models, with a 20% reduction in FLOPs . These findings underscore the potential of our approach in advancing real‐time traffic sign recognition technologies.
Zitieren
APA
Wu, Y., Chen, Z., & Lu, W. (2026). Enhanced Real‐Time Traffic Sign Recognition via Lightweight Neural Networks and Wavelet Transform. IET Intelligent Transport Systems, 20(1). https://doi.org/10.1049/itr2.70286
Vancouver
Wu Y, Chen Z, Lu W. Enhanced Real‐Time Traffic Sign Recognition via Lightweight Neural Networks and Wavelet Transform. IET Intelligent Transport Systems. 2026;20(1). doi:10.1049/itr2.70286
BibTeX
@article{Wu2026,
author = {Wu, Yi‐Chao and Chen, Zu‐Song and Lu, Wei‐Jin},
title = {Enhanced Real‐Time Traffic Sign Recognition via Lightweight Neural Networks and Wavelet Transform},
journal = {IET Intelligent Transport Systems},
year = {2026},
volume = {20},
number = {1},
doi = {10.1049/itr2.70286}
}
EUVIMED