A lightweight underwater fish image semantic segmentation model based on U‐Net

Abstract Semantic segmentation of underwater fish images is vital for monitoring fish stocks, assessing marine resources, and sustaining fisheries. To tackle challenges such as low segmentation accuracy, inadequate real‐time performance, and imprecise location segmentation in current methods, a nove...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhenkai Zhang, Wanghua Li, Boon‐Chong Seet
Format: Article
Language:English
Published: Wiley 2024-10-01
Series:IET Image Processing
Subjects:
Online Access:https://doi.org/10.1049/ipr2.13161
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850200406490087424
author Zhenkai Zhang
Wanghua Li
Boon‐Chong Seet
author_facet Zhenkai Zhang
Wanghua Li
Boon‐Chong Seet
author_sort Zhenkai Zhang
collection DOAJ
description Abstract Semantic segmentation of underwater fish images is vital for monitoring fish stocks, assessing marine resources, and sustaining fisheries. To tackle challenges such as low segmentation accuracy, inadequate real‐time performance, and imprecise location segmentation in current methods, a novel lightweight U‐Net model is proposed. The proposed model acquires more segmentation details by applying a multiple‐input approach at the first four encoder levels. To achieve both lightweight and high accuracy, a multi‐scale residual structure (MRS) module is proposed to reduce parameters and compensate for the accuracy loss caused by the reduction of channels. To improve segmentation accuracy, a multi‐scale skip connection (MSC) structure is further proposed, and the convolution block attention mechanism (CBAM) is introduced at the end of each decoder level for weight adjustment. Experimental results demonstrate a notable reduction in model volume, parameters, and floating‐point operations by 94.20%, 94.39%, and 51.52% respectively, compared to the original model. The proposed model achieves a high mean intersection over union (mIOU) of 94.44%, mean pixel accuracy (mPA) of 97.03%, and a frame rate of 43.62 frames per second (FPS). With its high precision and minimal parameters, the model strikes a balance between accuracy and speed, making it particularly suitable for underwater image segmentation.
format Article
id doaj-art-68190b95e42e4d368e1e963642bb0eda
institution OA Journals
issn 1751-9659
1751-9667
language English
publishDate 2024-10-01
publisher Wiley
record_format Article
series IET Image Processing
spelling doaj-art-68190b95e42e4d368e1e963642bb0eda2025-08-20T02:12:20ZengWileyIET Image Processing1751-96591751-96672024-10-0118123143315510.1049/ipr2.13161A lightweight underwater fish image semantic segmentation model based on U‐NetZhenkai Zhang0Wanghua Li1Boon‐Chong Seet2Ocean College Jiangsu University of Science and Technology Zhenjiang People's Republic of ChinaOcean College Jiangsu University of Science and Technology Zhenjiang People's Republic of ChinaDepartment of Electrical and Electronic Engineering Auckland University of Technology Auckland New ZealandAbstract Semantic segmentation of underwater fish images is vital for monitoring fish stocks, assessing marine resources, and sustaining fisheries. To tackle challenges such as low segmentation accuracy, inadequate real‐time performance, and imprecise location segmentation in current methods, a novel lightweight U‐Net model is proposed. The proposed model acquires more segmentation details by applying a multiple‐input approach at the first four encoder levels. To achieve both lightweight and high accuracy, a multi‐scale residual structure (MRS) module is proposed to reduce parameters and compensate for the accuracy loss caused by the reduction of channels. To improve segmentation accuracy, a multi‐scale skip connection (MSC) structure is further proposed, and the convolution block attention mechanism (CBAM) is introduced at the end of each decoder level for weight adjustment. Experimental results demonstrate a notable reduction in model volume, parameters, and floating‐point operations by 94.20%, 94.39%, and 51.52% respectively, compared to the original model. The proposed model achieves a high mean intersection over union (mIOU) of 94.44%, mean pixel accuracy (mPA) of 97.03%, and a frame rate of 43.62 frames per second (FPS). With its high precision and minimal parameters, the model strikes a balance between accuracy and speed, making it particularly suitable for underwater image segmentation.https://doi.org/10.1049/ipr2.13161computer visionconvolutional neural netsimage segmentationoceanographic techniques
spellingShingle Zhenkai Zhang
Wanghua Li
Boon‐Chong Seet
A lightweight underwater fish image semantic segmentation model based on U‐Net
IET Image Processing
computer vision
convolutional neural nets
image segmentation
oceanographic techniques
title A lightweight underwater fish image semantic segmentation model based on U‐Net
title_full A lightweight underwater fish image semantic segmentation model based on U‐Net
title_fullStr A lightweight underwater fish image semantic segmentation model based on U‐Net
title_full_unstemmed A lightweight underwater fish image semantic segmentation model based on U‐Net
title_short A lightweight underwater fish image semantic segmentation model based on U‐Net
title_sort lightweight underwater fish image semantic segmentation model based on u net
topic computer vision
convolutional neural nets
image segmentation
oceanographic techniques
url https://doi.org/10.1049/ipr2.13161
work_keys_str_mv AT zhenkaizhang alightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet
AT wanghuali alightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet
AT boonchongseet alightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet
AT zhenkaizhang lightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet
AT wanghuali lightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet
AT boonchongseet lightweightunderwaterfishimagesemanticsegmentationmodelbasedonunet