EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

WaveMamba-YOLO: Combining frequency awareness and state-space modeling for defect localization

Aping Ge, Yang Lv, Jun Huang iD

PLOS One · 2026

DOI: 10.1371/journal.pone.0344940 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, 2026. Verfasst von Aping Ge und 2 weiteren Autor:innen. Beteiligte Einrichtung: Faculty of Intelligent Manufacturing, Wuhu Vocational Technical University, Wuhu, Anhui, China. Hauptthemen laut MEDLINE: Steel, Models, Theoretical. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: the Key Nature Science Research Projects of Anhui Province, Wuhu Institute of Technology Quality Engineering Projects, Wuhu Institute of Technology Talent Engineering Projects.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Steel surface defect detection is critical for ensuring the reliability and safety of automotive manufacturing. However, existing methods often suffer from high computational cost, weak sensitivity to fine textures, and limited adaptability to diverse defect scales. To address these challenges, we propose WaveMamba-YOLO, a real-time detection framework that integrates frequency-domain enhancement with efficient state-space modeling. The architecture introduces three key modules: (1) CHDWT, which combines Haar wavelet decomposition and residual learning to preserve structural details during downsampling; (2) GLaM, a global-local-aware Mamba module that couples large-kernel convolution with state-space modeling to capture long-range dependencies at linear complexity; and (3) LWGA, a lightweight group attention mechanism that adaptively attends to micro-, regular-, medium-, and large-scale defects. Experiments on the Severstal Steel Defect and NEU-DET datasets demonstrate that WaveMamba-YOLO achieves superior performance, reaching 51.70% mAP@0.5 and 58.60% precision on Severstal and 77.70% mAP@0.5 on NEU-DET, consistently surpassing mainstream lightweight detectors. These results confirm the effectiveness of WaveMamba-YOLO in balancing detection accuracy and efficiency, highlighting its potential for real-time industrial inspection.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

SteelModels, TheoreticalDetection AlgorithmsAlgorithms

Substanzen

Steel

Autor:innen und Institutionen

  • Faculty of Intelligent Manufacturing, Wuhu Vocational Technical University, Wuhu, Anhui, China.
  • Yalong Intelligent Equipment Group Co., Ltd, Wenzhou, Zhejiang, China.
  • Faculty of Information Technology, City University of Malaysia, Kuala Lumpur, Malaysia.

ORCID-Profile: Jun Huang

Förderung

  • the Key Nature Science Research Projects of Anhui Province
  • Wuhu Institute of Technology Quality Engineering Projects
  • Wuhu Institute of Technology Talent Engineering Projects
  • Wuhu Institute of Technology Intelligent Manufacturing Equipment Technology Teaching Innovation Team Projects
  • the General Program for the Training of Outstanding Young Teachers of Anhui Province

Zitieren

APA

Ge, A., Lv, Y., & Huang, J. (2026). WaveMamba-YOLO: Combining frequency awareness and state-space modeling for defect localization. PLOS One. https://doi.org/10.1371/journal.pone.0344940

Vancouver

Ge A, Lv Y, Huang J. WaveMamba-YOLO: Combining frequency awareness and state-space modeling for defect localization. PLOS One. 2026. doi:10.1371/journal.pone.0344940

BibTeX
@article{Ge2026,
  author = {Ge, Aping and Lv, Yang and Huang, Jun},
  title = {WaveMamba-YOLO: Combining frequency awareness and state-space modeling for defect localization},
  journal = {PLOS One},
  year = {2026},
  doi = {10.1371/journal.pone.0344940}
}
Kostenlos starten →