This repo holds the code of "MASDF-Net: A multi-attention codec network with selective and dynamic fusion for skin lesion segmentation."
GPUs of memory>=4G shall be sufficient for this experiment.
-
Preparing necessary data:
- downloading ISIC training, validation and testing data from the official site, put the unzipped data in
./data
. - run
process.py
to preprocess all the data, which generatesdata_{train, val, test}.npy
andmask_{train, val, test}.npy
.
- downloading ISIC training, validation and testing data from the official site, put the unzipped data in
-
Testing:
- run
test_isic.py --ckpt_path='snapshots/XXXX.pth'
.
- run
-
Training:
- run
train_isic.py
; you may also want to change the default saving path or other hparams as well.
- run
We are very grateful for these excellent works PVT, Polyp-PVT, CCNet,TransFuse and PraNet, which have provided the basis for our framework.
Please drop an email to fujinghao@stu.scu.edu.cn