GA-HQS: MRI reconstruction via a generically accelerated unfolding approach
This work addresses compressed sensing MRI reconstruction, a domain-specific problem, with incremental improvements in method design.
The paper tackled MRI reconstruction by proposing GA-HQS, a deep unfolding network that incorporates second-order gradient information and pyramid attention modules, achieving superior performance on single-coil MRI acceleration tasks compared to previous methods.
Deep unfolding networks (DUNs) are the foremost methods in the realm of compressed sensing MRI, as they can employ learnable networks to facilitate interpretable forward-inference operators. However, several daunting issues still exist, including the heavy dependency on the first-order optimization algorithms, the insufficient information fusion mechanisms, and the limitation of capturing long-range relationships. To address the issues, we propose a Generically Accelerated Half-Quadratic Splitting (GA-HQS) algorithm that incorporates second-order gradient information and pyramid attention modules for the delicate fusion of inputs at the pixel level. Moreover, a multi-scale split transformer is also designed to enhance the global feature representation. Comprehensive experiments demonstrate that our method surpasses previous ones on single-coil MRI acceleration tasks.