Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture

Real-time estimation of fish biomass plays a crucial role in real-world fishery production, as it helps formulate feeding strategies and other management decisions. In this paper, a dense fish counting network called Swin-CSRNet is proposed. Specifically, the VGG16 layer in the front-end is replaced...

Full description

Saved in:
Bibliographic Details
Main Authors: Jintao Liu, Alfredo Tolón-Becerra, José Fernando Bienvenido-Barcena, Xinting Yang, Kaijie Zhu, Chao Zhou
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Journal of Marine Science and Engineering
Subjects:
Online Access:https://www.mdpi.com/2077-1312/12/10/1823
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850206169948225536
author Jintao Liu
Alfredo Tolón-Becerra
José Fernando Bienvenido-Barcena
Xinting Yang
Kaijie Zhu
Chao Zhou
author_facet Jintao Liu
Alfredo Tolón-Becerra
José Fernando Bienvenido-Barcena
Xinting Yang
Kaijie Zhu
Chao Zhou
author_sort Jintao Liu
collection DOAJ
description Real-time estimation of fish biomass plays a crucial role in real-world fishery production, as it helps formulate feeding strategies and other management decisions. In this paper, a dense fish counting network called Swin-CSRNet is proposed. Specifically, the VGG16 layer in the front-end is replaced with the Swin transformer to extract image features more efficiently. Additionally, a squeeze-and-excitation (SE) module is introduced to enhance feature representation by dynamically adjusting the importance of each channel through “squeeze” and “excitation”, making the extracted features more focused and effective. Finally, a multi-scale fusion (MSF) module is added after the back-end to fully utilize the multi-scale feature information, enhancing the model’s ability to capture multi-scale details. The experiment demonstrates that Swin-CSRNet achieved excellent results with MAE, RMSE, and MAPE and a correlation coefficient R<sup>2</sup> of 11.22, 15.32, 5.18%, and 0.954, respectively. Meanwhile, compared to the original network, the parameter size and computational complexity of Swin-CSRNet were reduced by 70.17% and 79.05%, respectively. Therefore, the proposed method not only counts the number of fish with higher speed and accuracy but also contributes to advancing the automation of aquaculture.
format Article
id doaj-art-adfd57659204446b8ea71ad61efb2ef2
institution OA Journals
issn 2077-1312
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Journal of Marine Science and Engineering
spelling doaj-art-adfd57659204446b8ea71ad61efb2ef22025-08-20T02:10:54ZengMDPI AGJournal of Marine Science and Engineering2077-13122024-10-011210182310.3390/jmse12101823Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in AquacultureJintao Liu0Alfredo Tolón-Becerra1José Fernando Bienvenido-Barcena2Xinting Yang3Kaijie Zhu4Chao Zhou5School of Engineering, University of Almeria, 04120 Almeria, SpainSchool of Engineering, University of Almeria, 04120 Almeria, SpainSchool of Engineering, University of Almeria, 04120 Almeria, SpainNational Engineering Research Center for Information Technology in Agriculture, Beijing 100097, ChinaNational Engineering Research Center for Information Technology in Agriculture, Beijing 100097, ChinaNational Engineering Research Center for Information Technology in Agriculture, Beijing 100097, ChinaReal-time estimation of fish biomass plays a crucial role in real-world fishery production, as it helps formulate feeding strategies and other management decisions. In this paper, a dense fish counting network called Swin-CSRNet is proposed. Specifically, the VGG16 layer in the front-end is replaced with the Swin transformer to extract image features more efficiently. Additionally, a squeeze-and-excitation (SE) module is introduced to enhance feature representation by dynamically adjusting the importance of each channel through “squeeze” and “excitation”, making the extracted features more focused and effective. Finally, a multi-scale fusion (MSF) module is added after the back-end to fully utilize the multi-scale feature information, enhancing the model’s ability to capture multi-scale details. The experiment demonstrates that Swin-CSRNet achieved excellent results with MAE, RMSE, and MAPE and a correlation coefficient R<sup>2</sup> of 11.22, 15.32, 5.18%, and 0.954, respectively. Meanwhile, compared to the original network, the parameter size and computational complexity of Swin-CSRNet were reduced by 70.17% and 79.05%, respectively. Therefore, the proposed method not only counts the number of fish with higher speed and accuracy but also contributes to advancing the automation of aquaculture.https://www.mdpi.com/2077-1312/12/10/1823aquaculturefish countingdensity mapSwin transformer
spellingShingle Jintao Liu
Alfredo Tolón-Becerra
José Fernando Bienvenido-Barcena
Xinting Yang
Kaijie Zhu
Chao Zhou
Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
Journal of Marine Science and Engineering
aquaculture
fish counting
density map
Swin transformer
title Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
title_full Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
title_fullStr Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
title_full_unstemmed Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
title_short Hybrid Swin-CSRNet: A Novel and Efficient Fish Counting Network in Aquaculture
title_sort hybrid swin csrnet a novel and efficient fish counting network in aquaculture
topic aquaculture
fish counting
density map
Swin transformer
url https://www.mdpi.com/2077-1312/12/10/1823
work_keys_str_mv AT jintaoliu hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture
AT alfredotolonbecerra hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture
AT josefernandobienvenidobarcena hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture
AT xintingyang hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture
AT kaijiezhu hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture
AT chaozhou hybridswincsrnetanovelandefficientfishcountingnetworkinaquaculture