[go: up one dir, main page]
More Web Proxy on the site http://driver.im/ skip to main content
10.1145/3652628.3652705acmotherconferencesArticle/Chapter ViewAbstractPublication PagesicaiceConference Proceedingsconference-collections
research-article

A Smoke Segmentation Method Based on Seaformer

Published: 23 May 2024 Publication History

Abstract

Aiming at the problem that the smoke is irregular in shape, translucent and blurred in boundary, which leads to the difficulty of smoke segmentation, this paper applies a dual-branch efficient and lightweight semantic segmentation algorithm with attention mechanism and fusion of context information and spatial information to smoke semantic segmentation. The network structure mainly uses the MobileNetV2 network to extract the visual smoke features, and combines the Seaformer (squeeze-enhanced Axial Transformer ) attention mechanism. Finally, a lightweight smoke segmentation head is connected to achieve the segmentation of smoke texture information when the size of parameters is as small as possible. According to the characteristics of deep learning models that require a large number of training sets, the quantity and quality of training sets always directly affect the quality of the model. Therefore, this paper uses a random real-time synthesis of smoke to increase smoke samples. The segmentation threshold is set to 0.2, and the loss function is changed to MSE ( Mean squarre error ) to adapt to the binary classification task of smoke image segmentation. Finally, the comparison experiments with some classical semantic segmentation models indicate that the model in this paper achieves the best mIoU(Mean Intersection over Union) scores on synthetic datasets. And it also exhibits best results in real smoke testing data.

References

[1]
YU J,LIN Z,YANG J,et al. 2018. Generative image inpainting with contextual attention[C]//Proceedings of the IEEE Conference on Computer Vi sion and Pattern Recognition.IEEE, 2018: 5505-5514.
[2]
Su T, Wang Y, Deng Q Y 2023. Improved foggy pedestrian and vehicle detection algorithm based on YOLOv5 [ J / OL ].System Simulation Journal : 1-11. 2023. 11. 12. https://doi.org/10.16182/j.issn1004731x.joss.23-0725.
[3]
Jin W Z, Lv W T, Guo Q 2022. Fabric image classification algorithm based on improved 3E-LDA [ J ].Modern textile technology.
[4]
Tang H, Wang M, Yu Q S, etc. 2023. Root image segmentation method based on improved UNet and transfer learning [ J ].Smart agriculture ( Chinese and English ).
[5]
Tian X, Wang L, Ding Q. 2019. An overview of image semantic segmentation methods based on deep learning [ J ]. JOURNAL OF SOFTWARE, 2019,30 ( 02 ).
[6]
LONG J, SHELHAMER E, DARRELL T. 2015. Fully convolutional networks for semantic segmentation [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston: IEEE, 2015: 3431–3440.
[7]
Chen J Z, Wang Z J, Chen H H, etc. 2016. Video dynamic smoke detection based on cascaded convolutional neural network [ J ].Journal of University of Electronic Science and Technology of China. 2016,45(6):992-996.
[8]
Hu Y,Zhan J,Zhou G,et al. 2022. Fast forest fire smoke detection using MVMNet[J]. Knowledge-Based Systems,2022,241:108219.
[9]
Yuan F N,Zhang L,Wan B Y,et al. 2019. Convolutional neural networks based on multi-scaleadditive merging layers for visual smoke recognition[J].Machine Vision and Applications,2019,30(2):345-358.
[10]
LIU Z,LIN Y,CAO Y,et al. 2021. Swin Transformer:Hierarchical Vision Transformer Using Shifted Windows[C]//Proceedings of the IEEE/CVF Int ernational Conference on Computer Vision.Montreal,QC,Canada:IEEE,2021: 10012-10022.
[11]
Badrinarayanan V,Kendall A,Cipolla R. 2017. SegNet:A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation.IEEE Transactions on Pattern Analysis Machine Intelligence, 2017, 39(12):2841-2495.
[12]
Wan Q,Huang Z,Lu J,et al. 2023. Seaformer:Squeeze-enhanced Axial Transformer for Mobile Semantic Segmentation[J].arXiv preprint arXiv:2 301.13156,2023.

Index Terms

  1. A Smoke Segmentation Method Based on Seaformer

    Recommendations

    Comments

    Please enable JavaScript to view thecomments powered by Disqus.

    Information & Contributors

    Information

    Published In

    cover image ACM Other conferences
    ICAICE '23: Proceedings of the 4th International Conference on Artificial Intelligence and Computer Engineering
    November 2023
    1263 pages
    ISBN:9798400708831
    DOI:10.1145/3652628
    Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].

    Publisher

    Association for Computing Machinery

    New York, NY, United States

    Publication History

    Published: 23 May 2024

    Permissions

    Request permissions for this article.

    Check for updates

    Qualifiers

    • Research-article
    • Research
    • Refereed limited

    Conference

    ICAICE 2023

    Contributors

    Other Metrics

    Bibliometrics & Citations

    Bibliometrics

    Article Metrics

    • 0
      Total Citations
    • 5
      Total Downloads
    • Downloads (Last 12 months)5
    • Downloads (Last 6 weeks)0
    Reflects downloads up to 01 Jan 2025

    Other Metrics

    Citations

    View Options

    Login options

    View options

    PDF

    View or Download as a PDF file.

    PDF

    eReader

    View online with eReader.

    eReader

    HTML Format

    View this article in HTML Format.

    HTML Format

    Media

    Figures

    Other

    Tables

    Share

    Share

    Share this Publication link

    Share on social media