Inter-scale Context Fusion

The official code for "Inter-Scale Dependency Modeling for Skin Lesion Segmentation with Transformer-based Networks".

Updates

• 1 May., 2023 : Initial release.
• 28 Apr., 2023: Accepted.
• 7 Apr., 2023: Submitted to MIDL 2023 [Under Review].

Citation

@inproceedings{eskandari2023interscale,
    title={Inter-Scale Dependency Modeling for Skin Lesion Segmentation with Transformer-based Networks},
    author={Sania Eskandari and Janet Lumpp},
    booktitle={Medical Imaging with Deep Learning, short paper track},
    year={2023},
    url={https://openreview.net/forum?id=JExQEfV5um}
}

---

Setting up and Training

• In order to run the code and experiments, you need to first install the dependencies and then download and move the data to the right directory.

• For ISIC 2017-18 datasets, we used the ISIC Challenge datasets link.

• Run the following code to install the Requirements.

  pip install -r requirements.txt

---

Model weights

You can download the learned weights of the DAEFormer in the following table.

Task	Dataset	Learned weights
Skin Lesion Segmentation	ISIC 2017	ISCF
Skin Lesion Segmentation	ISIC 2018	ISCF

---

References

This repo heavily built on the following repos.

• Swin-Unet
• SegFormer
• MISSFormer

---