OCTCube Models Release

This repository hosts the official release of OCTCube-M: A 3D multimodal optical coherence tomography foundation model for retinal and systemic diseases with cross-cohort and cross-device validation: [Github repo] | [Arxiv].

We release OCTCube, OCTCube-IR, and a Multi-task Classification Model for 8 retinal diseases. These models are designed for multi-modal OCT analysis and support image classification tasks in ophthalmology.

🔹 Available Models

Model Name	Description
OCTCube	A foundation model trained on OCT data for learning shared embeddings.
OCTCube-IR	An infrared-enhanced version of OCTCube that integrates IR modality.
Multi-task Classification Model	A model trained for multi-disease classification of 8 retinal diseases.

📜 Model Details

OCTCube: Designed for learning joint representations of OCT images.
OCTCube-IR: Extends OCTCube by incorporating IR modality for improved alignment.
Multi-task Classification Model: A specialized classification model for 8 retinal diseases, leveraging multi-task learning.

📂 Example Data

oct_examples/: Example OCT data volumes provided for trying our inference notebook! Includes:
- AMD.dcm
- DR.dcm
- POAG.dcm
- DME.dcm
- RNV.dcm
- VD.dcm
- ERM.dcm
- CRAO/CRVO.dcm

🔧 Usage

To use the models in PyTorch, please check the inference noteboook.

📑 Citation

If you use these models in your research, please cite:

@article{liu2024octcube,
  title={OCTCube: a 3D foundation model for optical coherence tomography that improves cross-dataset, cross-disease, cross-device and cross-modality analysis},
  author={Liu, Zixuan and Xu, Hanwen and Woicik, Addie and Shapiro, Linda G and Blazes, Marian and Wu, Yue and Lee, Cecilia S and Lee, Aaron Y and Wang, Sheng},
  journal={arXiv preprint arXiv:2408.11227},
  year={2024}
}

💬 Contact

For any questions or feedback, feel free to open an issue or contact us directly ([email protected]).

🚀 Stay tuned for updates and improvements!

zucksliu
/

OCTCubeM