Abstract:The recognition of underwater acoustic targets plays a crucial role in marine vessel monitoring. However, traditional underwater target recognition models suffer from limitations, including low recognition accuracy and slow prediction speed. To address these challenges, this article introduces a novel approach called the Multi-Gradient Flow Global Feature Enhancement Network (MGFGNet) for automatic recognition of underwater acoustic targets. Firstly, a new spectrogram feature fusion scheme is presented, effectively capturing both the physical and brain-inspired features of the acoustic signal. This fusion technique enhances the representation of underwater acoustic data, resulting in more accurate recognition results. Moreover, MGFGNet utilizes the multi-gradient flow network and incorporates a multi-dimensional feature enhancement technique to achieve fast and precise end-to-end recognition. Finally, a loss function is introduced to mitigate the influence of unbalanced data sets on model recognition performance using Taylor series. This further enhances model recognition performance. Experimental evaluations were conducted on the DeepShip dataset to assess the performance of our proposed method. The results demonstrate the superiority of MGFGNet, achieving a recognition rate of 99.1%, which significantly surpasses conventional methods. Furthermore, MGFGNet exhibits improved efficiency compared to the widely used ResNet18 model, reducing the parameter count by 51.28% and enhancing prediction speed by 33.9%. Additionally, we evaluated the generalization capability of our model using the ShipsEar dataset, where MGFGNet achieves a recognition rate of 99.5%, indicating its superior performance when applied to unbalanced data. The promising results obtained in this study highlight the potential of MGFGNet in practical applications.

PLGAT: Underwater Plectropomus leopardus Recognition Using Global Attention Mechanism and Transfer Learning

YoloXT: A Object Detection Algorithm for Marine Benthos

MGFGNet: an Automatic Underwater Acoustic Target Recognition Method Based on the Multi-Gradient Flow Global Feature Enhancement Network

A Novel Underwater Image Enhancement and Improved Underwater Biological Detection Pipeline

Underwater Object Detection Network with Dilated Deformable Convolutions and Dual-Branch Occlusion Attention Mechanism

A Novel Underwater Image Enhancement Algorithm and an Improved Underwater Biological Detection Pipeline

A new non-invasive tagging method for leopard coral grouper (Plectropomus leopardus) using deep convolutional neural networks with PDE-based image decomposition

TransCODNet: Underwater Transparently Camouflaged Object Detection Via RGB and Event Frames Collaboration

Coastal Fisheries Resource Monitoring Through A Deep Learning-Based Underwater Video Analysis

PL-Net: towards deep learning-based localization for underwater terrain

Fish Recognition in the Underwater Environment Using an Improved ArcFace Loss for Precision Aquaculture

Probabilistic Model-Based Active Learning with Attention Mechanism for Fish Species Recognition

Adaptive Learning Attention Network for Underwater Image Enhancement

Underwater Species Detection using Channel Sharpening Attention

An Attention Fusion Enhanced Underwater Object Detection Algorithm

Instance Segmentation of Underwater Images by Using Deep Learning

A New Dataset, Poisson GAN and AquaNet for Underwater Object Grabbing

Self-Supervised Marine Organism Detection from Underwater Images

Integrate MSRCR and Mask R-CNN to Recognize Underwater Creatures on Small Sample Datasets

DeepFish: Accurate Underwater Live Fish Recognition with a Deep Architecture.

Underwater Fish Image Recognition Based on Detection Transformer