中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 深海科学与工程研究所
  3. 中国科学院深海科学与工程研究所
  4. 深海工程技术部
  5. 深海视频技术研究室
FishFocusNet: An improved method based on YOLOv8 for underwater tropical fish identification

文献类型：期刊论文

作者Lu, Zhaoxuan3; Zhu, Xiaolong2; Guo, Haitao2; Xie, Xingang3; Chen, Xiangzi2; Quan, Xiangqian1
刊名IET IMAGE PROCESSING
出版日期2024-08-07
页码16
关键词convolutional neural nets object detection object recognition
ISSN号1751-9659
DOI10.1049/ipr2.13202
英文摘要Accurately identifying tropical fish serves as a crucial indicator, offering an insight into the state of marine biodiversity and the condition of coral reef ecosystems. However, the current detection networks are prone to omission and misidentification due to occlusion between fish and the complex underwater environment. This paper proposes a modified approach named FishFocusNet, in which alterable kernel convolution modules, asymptotic feature pyramid network (AFPN), and Shape-IoU are integrated into YOLOv8. To extract a more comprehensive set of fish features, AKConv modules with arbitrary kernel sizes are proposed to take the place of the conventional fixed-shaped kernels in the backbone for downsampling. AFPN is adopted as the feature integration structure in the neck, which enhances feature fusion and adaptive spatial fusion between non-adjacent layers. In the detector head, Shape-IoU is employed to achieve precise localization of fish targets. The superiorities of these modifications are proved by ablation experiments and comparative experiments. The experimental results show that the optimized approach obtained an mAP of 81.8% accompanied by 2.4 MB parameters and 3.6 GB FLOPS. Meanwhile, compared with more complicated models of similar scale, the proposed method can enhance recognition accuracy to 84.2% and significantly reduce computational costs. An improved deep learning network FishFocusNet for underwater tropical fish identification is proposed. It was based on the latest YOLOv8 model, and it employed alterable kernel convolution modules, asymptotic feature pyramid network, and Shape-IoU to enhance the identification ability. This method is able to provide potential technical support for understanding marine biodiversity and the conservation of marine resources. image
资助项目Hainan Provincial Natural Science Foundation of China ; Science and Technology Innovation Special Project of Sanya in China[2022KJCX83] ; Scientific Research Foundation of Hainan Tropical Ocean University[RHDRC202205] ; South China Sea Institute of Oceanology, Chinese Academy of Sciences
WOS研究方向Computer Science ; Engineering ; Imaging Science & Photographic Technology
语种英语
WOS记录号WOS:001284849100001
出版者WILEY
资助机构Hainan Provincial Natural Science Foundation of China ; Science and Technology Innovation Special Project of Sanya in China ; Scientific Research Foundation of Hainan Tropical Ocean University ; South China Sea Institute of Oceanology, Chinese Academy of Sciences
源URL[http://ir.idsse.ac.cn/handle/183446/11475]  
专题深海工程技术部_深海视频技术研究室
通讯作者Zhu, Xiaolong
作者单位1.Chinese Acad Sci, Inst Deep Sea Sci & Engn, Sanya, Peoples R China
2.Hainan Trop Ocean Univ, Coll Marine Sci & Technol, Sanya 572022, Peoples R China
3.Hainan Trop Ocean Univ, Coll Marine Informat Engn, Sanya, Peoples R China
推荐引用方式
GB/T 7714		 Lu, Zhaoxuan,Zhu, Xiaolong,Guo, Haitao,et al. FishFocusNet: An improved method based on YOLOv8 for underwater tropical fish identification[J]. IET IMAGE PROCESSING,2024:16.
APA Lu, Zhaoxuan,Zhu, Xiaolong,Guo, Haitao,Xie, Xingang,Chen, Xiangzi,&Quan, Xiangqian.(2024).FishFocusNet: An improved method based on YOLOv8 for underwater tropical fish identification.IET IMAGE PROCESSING,16.
MLA Lu, Zhaoxuan,et al."FishFocusNet: An improved method based on YOLOv8 for underwater tropical fish identification".IET IMAGE PROCESSING (2024):16.

入库方式： OAI收割

来源：深海科学与工程研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。