This repository contains the codes corresponding to our MICCAI 2022 paper:

Mayur Mallya, Ghassan Hamarneh, "Deep Multimodal Guidance for Medical Image Classification", International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2022.

If you use the codes, please cite our paper. The bibtex entry is:

@inproceedings{mallya_2022_deepguide,
  title={Deep Multimodal Guidance for Medical Image Classification},
  author={Mallya, Mayur and Hamarneh, Ghassan},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year={2022},
  organization={Springer}
}

The proposed multimodal guidance strategy works as follows: (a) we first train the modality-specific classifiers C_I and C_S for both inferior and superior modalities, (b) next we train the guidance model G, followed by the guided inferior modality models G(I) and G(I)+I as in (c) and (d) respectively.

The codes for the analysis of brain tumor images of the RadPath dataset and skin lesion images of the Derm7pt dataset can be found in the respective folders.

If you find any issues in the repository and/or have suggestions on improving the code, feel free to raise an issue. Reach out to me via email at mmallya@sfu.ca for further collaborations and discussions.