A Self-supervised Diffusion Bridge for MRI Reconstruction
This addresses a limitation in medical imaging reconstruction for settings where reference images are unavailable, though it appears incremental as it builds on existing diffusion bridge frameworks.
The paper tackles the problem of training diffusion bridges for MRI reconstruction without requiring high-quality reference images, proposing SelfDB as a self-supervised method that achieves superior performance compared to denoising diffusion models in compressed sensing MRI.
Diffusion bridges (DBs) are a class of diffusion models that enable faster sampling by interpolating between two paired image distributions. Training traditional DBs for image reconstruction requires high-quality reference images, which limits their applicability to settings where such references are unavailable. We propose SelfDB as a novel self-supervised method for training DBs directly on available noisy measurements without any high-quality reference images. SelfDB formulates the diffusion process by further sub-sampling the available measurements two additional times and training a neural network to reverse the corresponding degradation process by using the available measurements as the training targets. We validate SelfDB on compressed sensing MRI, showing its superior performance compared to the denoising diffusion models.