MASNet: A Robust Deep Marine Animal Segmentation Network(Paper)

The Pytorch Implementation of ''MASNet: A Robust Deep Marine Animal Segmentation Network''

Introduction

In this project, we use Ubuntu 16.04.5, Python 3.7, Pytorch 1.7.1 and two NVIDIA RTX 2080Ti GPU.

Running

Testing

Download the pretrained model pre-trained model.

Check the model and image pathes in config.yaml and scripts/test.py, then run:

python test.py

Training

To train the model, you need to first prepare our RMAS dataset, or MAS3K dataset MAS3K dataset.

Check the dataset path in config.yaml, and then run:

python train.py

Citation

If you find MASNet is useful in your research, please cite our paper:

@ARTICLE{10113781,
  author={Fu, Zhenqi and Chen, Ruizhe and Huang, Yue and Cheng, En and Ding, Xinghao and Ma, Kai-Kuang},
  journal={IEEE Journal of Oceanic Engineering}, 
  title={MASNet: A Robust Deep Marine Animal Segmentation Network}, 
  year={2023},
  pages={1-12},
  doi={10.1109/JOE.2023.3252760}}