EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Efficient Smoke Segmentation Using Multiscale Convolutions and Multiview Attention Mechanisms

Xuesong Liu, Emmett J. Ientilucci

Electronics · 27.06.2025 · 14 (13) · 2593

DOI: 10.3390/electronics14132593 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 27.06.2025. Verfasst von Xuesong Liu und 1 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Austin Powder.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Efficient segmentation of smoke plumes is crucial for environmental monitoring and industrial safety. Existing models often face high computational demands and limited adaptability to diverse smoke appearances. To address these issues, we propose SmokeNet, a deep learning architecture integrating multiscale convolutions, multiview linear attention, and layer-specific loss functions. Specifically, multiscale convolutions capture diverse smoke shapes by employing varying kernel sizes optimized for different plume orientations. Subsequently, multiview linear attention emphasizes spatial and channel-wise features relevant to smoke segmentation tasks. Additionally, layer-specific loss functions promote consistent feature refinement across network layers, facilitating accurate and robust segmentation. SmokeNet achieves a segmentation accuracy of 72.74% mean Intersection over Union (mIoU) on our newly introduced quarry blast smoke dataset and maintains comparable performance on three benchmark smoke datasets, reaching up to 76.45% mIoU on the Smoke100k dataset. With a computational complexity of only 0.34 M parameters and 0.07 Giga Floating Point Operations (GFLOPs), SmokeNet is suitable for real-time applications. Evaluations conducted across these datasets demonstrate SmokeNet’s effectiveness and versatility in handling complex real-world scenarios.

Förderung

  • Austin Powder

Zitieren

APA

Liu, X., & Ientilucci, E. J. (2025). Efficient Smoke Segmentation Using Multiscale Convolutions and Multiview Attention Mechanisms. Electronics, 14(13), 2593. https://doi.org/10.3390/electronics14132593

Vancouver

Liu X, Ientilucci EJ. Efficient Smoke Segmentation Using Multiscale Convolutions and Multiview Attention Mechanisms. Electronics. 2025;14(13):2593. doi:10.3390/electronics14132593

BibTeX
@article{Liu2025,
  author = {Liu, Xuesong and Ientilucci, Emmett J.},
  title = {Efficient Smoke Segmentation Using Multiscale Convolutions and Multiview Attention Mechanisms},
  journal = {Electronics},
  year = {2025},
  volume = {14},
  number = {13},
  pages = {2593},
  doi = {10.3390/electronics14132593}
}
Kostenlos starten →