A Patch-Based Hankel Low-Rank Method for MRI Reconstruction

Motivation: Hankel low-rank reconstruction methods have shown the ability to produce high-quality reconstructions with lower reconstruction error and robustness to sampling patterns. However, their high computational complexity poses a significant challenge, particularly in high-dimensional imaging scenarios, limiting their applicability in situations requiring fast processing. Goal(s): We aim at accelerating Hankel low-rank reconstruction without sacrificing reconstruction quality. Approach: In this work, a patch-based Hankel low rank method is proposed by utilizing the low-rankness of a series of k-space patches. Results: Experimental results demonstrate our proposed scheme achieves approximately 4$$$\times$$$ acceleration compared to the traditional Hankel low-rank methods while maintaining comparable reconstruction errors. Impact: The proposed scheme has the flexibility to potentially improve all Hankel low-rank methods across various applications, providing fast and high-quality MRI reconstruction.