https://www.sciencedirect.com/science/article/pii/S1361841521003388
We cropped the ISIC 2018 dataset to 224*320 and saved it in npy format, which can be downloaded from Baidu web disk.
link: https://pan.baidu.com/s/1bIVUdzYG_7tuwalbI4Y8Ww
password: c36c
Place the downloaded npy files in the "data" directory and unzip them. The decompression format is as follows:
/data/ISIC2018_npy_all_224_320/image/
ISIC_0000000.npy
ISIC_0000001.npy
...
ISIC_0016072.npy
/data/ISIC2018_npy_all_224_320/label/
ISIC_0000000_segmentation.npy
ISIC_0000001_segmentation.npy
......
ISIC_0016072_segmentation.npy
Our program is easy to train and test, just need to run "main_train.py".
python main_train.py
| Networks | Para(M) | JI | Dice | ACC | Recall | Precision |
|---|---|---|---|---|---|---|
| FCN | 14.6 | 0.7866 | 0.8680 | 0.9504 | 0.8827 | 0.8784 |
| U-Net | 32.9 | 0.8169 | 0.8881 | 0.9568 | 0.8858 | 0.9131 |
| U-Net++ | 34.9 | 0.8187 | 0.8893 | 0.9568 | 0.8910 | 0.9098 |
| AttU-Net | 33.3 | 0.8199 | 0.8903 | 0.9577 | 0.8898 | 0.9126 |
| DeepLabv3+ | 37.9 | 0.8232 | 0.8926 | 0.9587 | 0.8974 | 0.9087 |
| DenseASPP | 33.7 | 0.8253 | 0.8935 | 0.9589 | 0.8950 | 0.9138 |
| CA-Net | 2.7 | 0.8041 | 0.8782 | 0.9525 | 0.8762 | 0.9072 |
| BCDU-Net | 28.8 | 0.8084 | 0.8833 | 0.9548 | 0.8913 | 0.8968 |
| Focus-Alpha | 26.4 | 0.8192 | 0.8893 | 0.9584 | 0.9157 | 0.8860 |
| DO-Net | 24.7 | 0.8261 | 0.8948 | 0.9578 | 0.9036 | 0.9059 |
| CE-Net | 29.0 | 0.8282 | 0.8959 | 0.9597 | 0.9054 | 0.9067 |
| CPF-Net | 43.3 | 0.8292 | 0.8963 | 0.9602 | 0.9062 | 0.9071 |
| Ms RED (our） | 3.8 | 0.8345 | 0.8999 | 0.9619 | 0.9049 | 0.9147 |
@article{dai2022ms,
title={Ms RED: A novel multi-scale residual encoding and decoding network for skin lesion segmentation},
author={Dai, Duwei and Dong, Caixia and Xu, Songhua and Yan, Qingsen and Li, Zongfang and Zhang, Chunyan and Luo, Nana},
journal={Medical Image Analysis},
volume={75},
pages={102293},
year={2022},
publisher={Elsevier}
}