An ISBI 2024 tutorial.

News

• 2024-05-28: The tutorial slides are publicly available here: DiMEDIA!
• 2024-05-28: The demonstration code is public on GitHub: ISBI 2024 DiMEDIA MONAI Tutorial!
• 2024-04-16: The tutorial will happen on 28 May 2024 - Tuesday afternoon (PM).
• 2024-01-09: The tutorial DiMEDIA: Diffusion Models in Medical Imaging and Analysis has been ACCEPTED for ISBI 2024!

Outline

Generative models have advanced rapidly in recent years, enabling the creation of high-quality synthetic data across images, volumes, and other complex medical data types. Among these models, diffusion methods have emerged as a powerful framework for image generation, reconstruction, denoising, segmentation, anomaly detection, and related tasks in medical imaging.

This tutorial presents an overview of generative modelling with a focus on diffusion models, including core theory, practical training considerations, conditioning strategies, and representative applications in medical image analysis. It builds on the MICCAI 2023 tutorial and extends it with updated material and demonstrations.

Figure from [1]

Tutorial Schedule

1. Part 1: Introduction (60 mins)

• What? Why? How?
• Denoising Diffusion Models
• Understanding and Intuition

2. Part 2: Advanced Topics (30 mins)

• Sampling Strategies
• Inference-time Conditioning
• Training-time Conditioning
• Trends in architecture and acceleration

3. Coffee break (30mins)
4. Applications in medical imaging (30 mins)

• Synthesis
• Reconstruction
• Segmentation
• Anomaly Detection
• Miscellaneous

5. Demonstration (30 mins)

• DEMO - MONAI Generative Models Coding tutorial on DDPM
• DEMO - MONAI Generative Models DDIM Inversion + Classifier-free guidance

6. Talk and Discussion (30 mins)

Learning Objectives

1. Understand the intuition and theory behind diffusion models
2. Present with demonstrations a software tool within MONAI (AI Toolkit for Healthcare Imaging) for applying diffusion models to medical imaging and image analysis
3. Understand the range of applications of diffusion models in medical imaging and image analysis
4. Appreciate current limitations of diffusion models

Organizing Team

• Mr. Yuyang Xue is a PhD student at the University of Edinburgh

• Ms. Nefeli Gkouti is a PhD student at Archimedes RU / Athena RC and at the National and Kapodistrian University of Athens.

• Dr. Julia Wolleb is a postdoctoral researcher at the Department of Biomedical Engineering at the University of Basel.

• Prof Sotirios A. Tsaftaris is the Canon Medical / Royal Academy of Engineering Research Chair in Healthcare AI with the University of Edinburgh, United Kingdom.

Some Resources

1. Kazerouni, Amirhossein, et al. “Diffusion models for medical image analysis: A comprehensive survey.” arXiv preprint arXiv:2211.07804 (2022).
2. https://github.com/heejkoo/Awesome-Diffusion-Models is a useful GitHub repository tracking recent diffusion-model papers.