Skip to main content
SpringerLink
Log in

Fish recognition in complex underwater scenes based on targeted sample transfer learning

  • Published:
Multimedia Tools and Applications Aims and scope Submit manuscript

Abstract

Fish population survey based on classification and recognition is an effective means to study water ecosystem. However, the identification of fish in the sea and other waters will be interfered by corals, reefs and other organisms. The great variety of fish also make it more difficult to distinguish. In order to improve the effect of fish recognition in the complex underwater environment, this paper proposes a method based on targeted sample transfer learning. The designed CNN is used to train the simple background fish data after background re-processing to obtain the pre-training model. By using transfer learning and interlayer fusion mechanism, the feature extraction layer of the pre-training model is frozen and fused with the new feature extraction layer in parallel, then, combined with pooling layer, a new feature extractor is formed, finally, connected to the classifier and output part to construct a new network, which is used to identify 10 kinds of fish with complex background. Compared with the original CNN model, the accuracy of the new network is improved by about 5%, reaching 91.33%. The experimental results show that the model can improve the ability of fish classification and recognition in complex underwater scenes, and can provide support for the study of fishery resources distribution.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11

Similar content being viewed by others

Data availability

All data and materials support our published claims and comply with field standards.

Code availability

All custom code support our published claims and comply with field standards.

References

  1. Aa A et al (2020) Deep learning-based cross-machine health identification method for vacuum pumps with domain adaptation. Procedia Manuf 48:1088–1093

    Article  Google Scholar 

  2. Albahli W, Albattah W (2020) Detection of coronavirus disease from X-ray images using deep learning and transfer learning algorithms. J X-Ray Sci Technol 28.5

  3. Arthington AH et al (2016) Fish conservation in freshwater and marine realms: status, threats and management. Aquat Conserv: Mar Freshw Ecosyst 26(5):838–857

    Article  Google Scholar 

  4. Moslem Azamfar, Xiang Li, Jay Lee (2020)Intelligent ball screw fault diagnosis using a deep domain adaptation methodology. Mechanism and Machine Theory,Volume 151,103932,ISSN 0094-114X. https://doi.org/10.1016/j.mechmachtheory.2020.103932

  5. Boom BJ, Huang PX, He J, Fisher RB (2012) Supporting ground-truth annotation of image datasets using clustering, in: 2012 21st International Conference on Pattern Recognition (ICPR), IEEE, Tsukuba, Japan, pp 1542–1545

  6. Chan A, Hodgson PA (2017) A systematic analysis of blast fishing in South-East Asia and possible solutions. 2017 IEEE Underwater Technology (UT). IEEE

  7. Chan A, Hodgson PA (2019) A belt transect fish abundance survey methodology using an underwater vehicle. 2019 IEEE Underwater Technology (UT). IEEE

  8. Chen W et al (2019) Fish classification based on deep convolutional neural network and transfer learning. Journal of Fuqing Branch of Fujian Normal University, No.5, Sum 156

  9. Dai Y, Guojun W, Li K-C(2018) Conceptual alignment deep neural networks. J Intell Fuzzy Syst, 1631–1642

  10. French B et al (2021) Comparing five methods for quantifying abundance and diversity of fish assemblages in seagrass habitat. Ecol Ind 124(4):107415

    Article  Google Scholar 

  11. Hodgson G (2001) Reef Check: The first step in community-based management. Bull. Mar. Sci., 69, pp. 861-868

  12. Huang Y, Wang Z (2020) Multi-granularity pruning for deep residual networks. J Intell Fuzzy Syst, 7403–7410. https://doi.org/10.3233/JIFS-200771

  13. Li J et al (2021) Deep neural network-based real time fish detection method in the scene of marine fishing supervision. J Intell Fuzzy Syst, 1–6. https://doi.org/10.3233/JIFS-189713

  14. Li J, Wu W, Xue D (2020) An intrusion detection method based on active transfer learning. Intell Data Anal, 363–383. https://doi.org/10.3233/IDA-194487

  15. Lifei, Wang et al (2018) Modeling marine pelagic fish species spatiotemporal distributions utilizing a maximum entropy approach. Fish Oceanogr 27(6):571–586

    Article  Google Scholar 

  16. Mostafa Mehdipour Ghazi, Berrin Yanikoglu, Erchan Aptoula (2017) Plant identification using deep neural networks via optimization of transfer learning parameters. Neurocomputing,Volume 235,Pages 228-235,ISSN 0925-2312. https://doi.org/10.1016/j.neucom.2017.01.018

  17. Macias-Garcia E et al (2021) Multi-stage deep learning perception system for mobile robots. Integr Comput Aided Eng, 191–205. https://doi.org/10.3233/ICA-200640

  18. Maruyama T et al (2018) Comparison of medical image classification accuracy among three machine learning methods. J X-Ray Sci Technol 26:1–93

    Article  Google Scholar 

  19. Masoudi B, Daneshvar S, Razavi SN (2021) Multi-modal neuroimaging feature fusion via 3D Convolutional neural network architecture for schizophrenia diagnosis. Intell Data Anal, 527–540. https://doi.org/10.3233/IDA-205113

  20. Pundhir S, Ghose U, Bisht U (2020) Assessment of effectiveness of data dependent activation method: MyAct. J Intell Fuzzy Syst, 665–677. https://doi.org/10.3233/JIFS-191618

  21. Qin H et al (2016) Deep fish: accurate underwater live fish recognition with a deep architecture. Neurocomputing 187:49–58

    Article  Google Scholar 

  22. Qiu C et al (2018) Transfer learning for small-scale fish image classification. 2018 OCEANS -. MTS/IEEE Kobe Techno-Ocean (OTO). IEEE

  23. Salman A et al (2016) Fish species classification in unconstrained underwater environments based on deep learning. Limnology and Oceanography: Methods 14. https://doi.org/10.1002/lom3.10113

  24. Yousaf, Waqas et al (2021) Patch-CNN: Deep learning for logo detection and brand recognition. J Intell Fuzzy Syst, 3849–3862. https://doi.org/10.3233/JIFS-190660

  25. Xi, Qiao et al (2019) fvUnderwater sea cucumber identification based on Principal Component Analysis and Support Vector Machin. Measurement. https://doi.org/10.1016/j.measurement.2018.10.039

  26. Yang F et al (2019) Quantification of hepatic steatosis in histologic images by deep learning method. J X-Ray Sci Technol, 1033–1045. https://doi.org/10.3233/XST-190570

  27. Zheng Z, Fu J, Lu C et al (2021) Research on rolling bearing fault diagnosis of small dataset based on a new optimal transfer learning network. Measurement. https://doi.org/10.1016/j.measurement.2021.109285

  28. M. D. Zeiler (2012) ADADELTA: An adaptive learning rate method. Computer ence, arXiv:1212.5701[cs.LG].

Download references

Author information

Authors and Affiliations

  1. Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, China

    Liangwei Jiang & Haiyan Quan

  2. Faculty of Civil Aviation and Aeronautics, Kunming University of Science and Technology, Kunming, China

    Tao Xie & Junbing Qian

Authors
  1. Liangwei Jiang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Haiyan Quan
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Tao Xie
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Junbing Qian
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Liangwei Jiang, Haiyan Quan, Tao Xie and Junbing Qian. The first draft of the manuscript was written by Liangwei Jiang and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Junbing Qian.

Ethics declarations

Conflicts of interest/Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Ethics approval

The paper complies with ethical standards.

Consent to participate

All authors give consent to participate.

Consent for publication

All authors give consent for publication.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jiang, L., Quan, H., Xie, T. et al. Fish recognition in complex underwater scenes based on targeted sample transfer learning. Multimed Tools Appl 81, 25303–25317 (2022). https://doi.org/10.1007/s11042-022-12525-8

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-022-12525-8

Keywords

Search

Navigation