Note: Due to the size of the datasets, the audio files were not uploaded to the repository. DeepShip dataset was used. Please follow the following instructions to ensure the code works. If the dataset is used, please cite the appropiate source (paper, repository, etc.) as mentioned on the webpage and provided here.
Please download the DeepShip dataset
and follow these instructions:
- Create a folder called
Deepship - Download the signals for each class in the
Deepshipfolder. - The structure of the
Deepshipfolder is as follows:
Deepship/
├── Cargo/
│ ├── Cargo1/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ ├── Cargo2/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ └── ...
├── Passenger/
│ ├── Passenger1/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ ├── Passenger2/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ └── ...
├── Tanker/
│ ├── Tanker1/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ ├── Tanker2/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ └── ...
├── Tug/
│ ├── Tug1/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ ├── Tug2/
│ │ ├── audio1.wav
│ │ ├── audio2.wav
│ │ └── ...
│ └── ...
└── ...
- The dataset structure after segmentation is shown in the figure below, along with the number of segments in each class.
[Note: The number of files for Cargo, Passengership, Tanker, and Tug is 109,191, 240, and 69, respectively after downloading from Google Drive.]
If you use the DeepShip dataset, please cite the following reference using the following entry.
Plain Text:
Irfan, M., Jiangbin, Z., Ali, S., Iqbal, M., Masood, Z., & Hamid, U. (2021). DeepShip: An underwater acoustic benchmark dataset and a separable convolution based autoencoder for classification. Expert Systems with Applications, 183, 115270.
BibTex:
@article{irfan2021deepship,
title={DeepShip: An underwater acoustic benchmark dataset and a separable convolution based autoencoder for classification},
author={Irfan, Muhammad and Jiangbin, Zheng and Ali, Shahid and Iqbal, Muhammad and Masood, Zafar and Hamid, Umar},
journal={Expert Systems with Applications},
volume={183},
pages={115270},
year={2021},
publisher={Elsevier}
}