Diffusion model for medical image denoising, reconstruction and translation

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Computer Vision

Background:

• Diffusion models are generative models that learn to reverse a noise-adding process.
• They have shown success in general image generation and various image processing tasks.
• Their application in medical imaging is emerging, showing potential for enhanced diagnostic accuracy and image quality.

Purpose of the Study:

• To provide a comprehensive overview of diffusion models in medical image analysis.
• To categorize diffusion models into denoising diffusion probabilistic models and score-based models.
• To summarize current applications and discuss future directions in medical imaging.

Main Methods:

• Review of theoretical foundations and fundamental concepts of diffusion models.
• Categorization into denoising diffusion probabilistic models and score-based models.
• Introduction to relevant datasets and evaluation metrics.

Main Results:

• Diffusion models show significant potential in medical image denoising, reconstruction, and translation.
• Detailed introductions to representative articles highlight current applications.
• Discussion of challenges and future research avenues.

Conclusions:

• Diffusion models offer a promising approach for advancing medical image analysis.
• Further research is needed to fully realize their potential in clinical settings.
• Continued development could lead to improved diagnostic tools and patient outcomes.