Motivation: With higher resolution, 5T MRI reveals anatomical details that 3T images cannot capture, enhancing diagnostic accuracy. However, due to the high cost and limited availability of 5T equipment, obtaining 5T SWI images is challenging. Goal(s): We propose an adaptive dual degradation diffusion model for synthesizing 5T SWI from 3T by learning image priors and degradation patterns. Approach: Our model utilizes two diffusion processes that model degradation diffusion and noise diffusion specifically to enhance control, detail, and robustness in synthesis. Results: Experimental results demonstrate the superiority and reliability of our method compared to other deep learning approaches. Impact: The method has shown effectiveness in enhancing SWI images, providing detailed results for brain microbleed evaluation and neurological assessments. These findings indicate that the proposed model has significant potential for advancing clinical applications in brain-related research.
An et al. (Tue,) studied this question.