Minimal Data Requirement for Realistic Endoscopic Image Generation with Stable Diffusion

This is an official repository for Minimal Data Requirement for Realistic Endoscopic Image Generation with Stable Diffusion.

Overview

This repository provides resources and workflows for generating realistic endoscopic images using Stable Diffusion. The methodology is divided into four key steps:

1. Training
   Train Stable Diffusion for endoscopic image generation.
2. Preprocessing
   Prepare raw simulation images for inference.
3. Inference
   Generate realistic endoscopic images using the trained model.
4. Evaluation
   Evaluate the quality and realism of the generated images.

Each step has a detailed README in its respective folder.

Resources

Access all resources, including datasets, pretrained models, and scripts: Dataset and Resources

Citation

If you use this work in your research, please cite it as follows:

@article{Kaleta2024,
  title={Minimal data requirement for realistic endoscopic image generation with Stable Diffusion},
  author={Kaleta, Joanna and Dall'Alba, Diego and Płotka, Szymon and Korzeniowski, Przemysław},
  journal={International Journal of Computer Assisted Radiology and Surgery},
  volume={19},
  number={3},
  pages={531--539},
  year={2024},
  doi={10.1007/s11548-023-03030-w},
  url={https://doi.org/10.1007/s11548-023-03030-w}
}