Magnetic resonance (MR) images collected in 2D clinical protocols typically have large inter-slice spacing, resulting in high in-plane resolution and reduced through-plane resolution. Super-resolution technique can enhance the through-plane resolution of MR images to facilitate downstream visualization and computer-aided diagnosis. However, most existing works train the super-resolution network at a fixed scaling factor, which is not friendly to clinical scenes of varying inter-slice spacing in MR scanning. Inspired by the recent progress in implicit neural representation, we propose a Spatial Attention-based Implicit Neural Representation (SA-INR) network for arbitrary reduction of MR inter-slice spacing. The SA-INR aims to represent an MR image as a continuous implicit function of 3D coordinates. In this way, the SA-INR can reconstruct the MR image with arbitrary inter-slice spacing by continuously sampling the coordinates in 3D space. In particular, a local-aware spatial attention operation is introduced to model nearby voxels and their affinity more accurately in a larger receptive field. Meanwhile, to improve the computational efficiency, a gradient-guided gating mask is proposed for applying the local-aware spatial attention to selected areas only. We evaluate our method on the public HCP-1200 dataset and the clinical knee MR dataset to demonstrate its superiority over other existing methods.
In clinical practice, some body parts (e.g., knee and ankle) typically only requires 2D protocols acquired LR MR images because of the shorter scan time. Therefore, paired LR/HR images are not popularly available to offer a full voxel-to-voxel supervision. We deploy our proposed model on a self-supervised reduction of MRI slice spacing framework to test our model's ability to apply to real-world clinical scenarios.
The results here are not cherry-picked (Please open with Chrome since we found bug with Safari)
Changing Slice Spacing from 4mm to 1mm. The following results are visualized by nearest up-sampling.