Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 14 min ago
Peer-reviewedCC BY 4.0PLOS One

FUSED-Net: Detecting traffic signs with limited data

  • Md. Atiqur Rahman1
  • Nahian Ibn Asad2
  • Md. Mushfiqul Haque3
  • Md. Bakhtiar Hasan
  • Sabbir Ahmed
  • Md. Hasanul Kabir
  1. 1Department of Computer Science and Engineering, Islamic University of Technology, Board Bazar, Gazipur, Dhaka, Bangladesh
  2. 2TigerIT Bangladesh Ltd., Banani, Dhaka, Bangladesh
  3. 3Department of Computer Science and Engineering, United International University, Badda, Dhaka, Bangladesh
Open at sourcePDFDOICite

Abstract

Automatic Traffic Sign Recognition is paramount in modern transportation systems. However, curating large-scale datasets for diverse traffic sign detection remains impractical. In this context, we present FUSED-Net, a novel approach that enhances Few-Shot Object Detection (FSOD) for traffic signs using limited data. FUSED-Net integrates Faster RCNN with Unfrozen Parameters, Pseudo-Support Sets, Embedding Normalization, and Domain Adaptation to improve detection accuracy. Unlike conventional methods, FUSED-Net keeps all parameters unfrozen during training, enabling it to learn effectively from limited samples. A Pseudo-Support Set is generated through data augmentation, enhancing performance by compensating for the scarcity of target domain data. Embedding Normalization reduces intra-class variance, standardizing feature representations. Domain Adaptation, achieved by pre-training on a diverse traffic sign dataset, improves model generalization. Experimental results on the BDTSD dataset demonstrate that FUSED-Net achieves 2.4×, 2.2×, 1.5×, and 1.3× improvements in mAP under 1-shot, 3-shot, 5-shot, and 10-shot scenarios, respectively, compared to state-of-the-art FSOD models. Additionally, FUSED-Net achieves superior performance on the cross-domain FSOD benchmark across multiple settings. The source code and the URLs to download the datasets are available at https://github.com/180041123-Atiq/FUSED-Net.

Related

  1. Asymmetric Reductive Three‐Component Annulation Reaction for Constructing Fused N‐Heterocycles With Contiguous Stereogenic Centers
    Guang‐peng Lu, Huan‐feng Jiang, Cheng‐gang Ci +2
    Angewandte Chemie International EditionPeer-reviewedLicence not stated1 citedSystems & synthetic biologyGeneral chemistryHAL
  2. Healthy Horse Vital Signs
    Sally DeNotta, Carissa Wickens
    EDISPeer-reviewedCC BY-NC-NDFull textGeneral biologyDOAJ
  3. Olfaction, cognition, and early signs of neurodegenerative disease in a community-based sample of older adults
    Erin Jacobsen, Yingjin Zhang, Mary Ganguli +5
    Alzheimer disease and associated disordersPeer-reviewedLicence not statedFull textNeurosciencePMC
  4. The self-sufficient Aspergillus fumigatus Δ9-fatty acid desaturase SdeA requires a fused cytochrome b5 domain and hinge-dependent self-association for virulence
    J. E. M. Campanella, G. Y. Filho, J. H. T. M. Fabri +6
    bioRxivPreprintAll rights reservedFull textMicrobiologybioRxiv
  5. Ocular sarcoidosis: from clinical signs to targeted interventions
    Thomas El Jammal, Robin Jacquot, Stéphane Abramowicz +5
    Current Opinion in ImmunologyPeer-reviewedCC BYGeneral biologyEpidemiology & public healthHAL

Cite

BibTeXRIS
BibTeX
@article{rahman2026fusednet,
  title = {{FUSED-Net: Detecting traffic signs with limited data}},
  author = {Md. Atiqur Rahman and Nahian Ibn Asad and Md. Mushfiqul Haque and Md. Bakhtiar Hasan and Sabbir Ahmed and Md. Hasanul Kabir},
  journal = {PLOS One},
  year = {2026},
  month = oct,
  doi = {10.1371/journal.pone.0359293},
  url = {https://doi.org/10.1371/journal.pone.0359293},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - FUSED-Net: Detecting traffic signs with limited data
AU  - Md. Atiqur Rahman
AU  - Nahian Ibn Asad
AU  - Md. Mushfiqul Haque
AU  - Md. Bakhtiar Hasan
AU  - Sabbir Ahmed
AU  - Md. Hasanul Kabir
PY  - 2026
DA  - 2026/10/07
JO  - PLOS One
DO  - 10.1371/journal.pone.0359293
UR  - https://doi.org/10.1371/journal.pone.0359293
AB  - Automatic Traffic Sign Recognition is paramount in modern transportation systems. However, curating large-scale datasets for diverse traffic sign detection remains impractical. In this context, we present FUSED-Net, a novel approach that enhances Few-Shot Object Detection (FSOD) for traffic signs using limited data. FUSED-Net integrates Faster RCNN with Unfrozen Parameters, Pseudo-Support Sets, Embedding Normalization, and Domain Adaptation to improve detection accuracy. Unlike conventional methods, FUSED-Net keeps all parameters unfrozen during training, enabling it to learn effectively from limited samples. A Pseudo-Support Set is generated through data augmentation, enhancing performance by compensating for the scarcity of target domain data. Embedding Normalization reduces intra-class variance, standardizing feature representations. Domain Adaptation, achieved by pre-training on a diverse traffic sign dataset, improves model generalization. Experimental results on the BDTSD dataset demonstrate that FUSED-Net achieves 2.4×, 2.2×, 1.5×, and 1.3× improvements in mAP under 1-shot, 3-shot, 5-shot, and 10-shot scenarios, respectively, compared to state-of-the-art FSOD models. Additionally, FUSED-Net achieves superior performance on the cross-domain FSOD benchmark across multiple settings. The source code and the URLs to download the datasets are available at https://github.com/180041123-Atiq/FUSED-Net.
SN  - 1932-6203
ER  -