SeLT: Sonar Echo Image Recognition for Small Targets using Lightweight Swin Transformer

被引：0

作者：

Xia, Sijia ^{[1
,2
]}

Hou, Mengyang ^{[1
,2
]}

Han, Yina ^{[1
,2
]}

Xiao, Ziyuan ^{[1
,2
]}

Guo, Zihao ^{[1
,2
]}

Liu, Qingyu ^{[3
]}

Ma, Yuanliang ^{[1
,2
]}

机构：

[1] Northwestern Polytech Univ, Sch Marine Sci & Technol, Xian, Peoples R China

[2] Shaanxi Key Lab Underwater Informat Technol, Xian, Peoples R China

[3] Naval Res Acad, Beijing, Peoples R China

来源：

OCEANS 2024 - SINGAPORE | 2024年

关键词：

Underwater acoustic target recognition; deep neural networks; lightweight network structure;

D O I：

10.1109/OCEANS51537.2024.10682372

中图分类号：

P75 [海洋工程];

学科分类号：

0814 ; 081505 ; 0824 ; 082401 ;

摘要：

Underwater sonar echo image data containing targets is relatively scarce, usually limiting the recognition performance of the model when employing a high-capacity (even state-of-the-art) network for recognition. To address this issue, we propose SeLT, a lightweight adaptation of the Swin-T, using lightweight feature extraction and feature coding modules. Specifically, we have reduced the stacking of Swin Transformer blocks and introduced a lightweight channel attention module to replace the MLP in each block. This eases the requirements for training data and computing resources, greatly accelerating the model training phase. Extensive experiments have demonstrated that, compared to the original Swin-T, our model achieves higher recognition performance (increasing 1.9% in terms of AUC value) with fewer parameters (reduced by 71%) and lower computational complexity (reduced by 73%).

引用

页数：5