Scale-Agnostic Super-Resolution in MRI using Feature-Based Coordinate Networks
This addresses the problem of achieving flexible super-resolution for MRI data, which is incremental as it adapts an existing coordinate network approach to a specific medical imaging task.
The authors tackled super-resolution in MRI by proposing a coordinate network decoder that is scale-gnostic, allowing training over continuous scales and querying at arbitrary resolutions, and they compared it to a convolutional decoder using quantitative metrics and a radiologist study, though no concrete numbers are provided in the abstract.
We propose using a coordinate network decoder for the task of super-resolution in MRI. The continuous signal representation of coordinate networks enables this approach to be scale-agnostic, i.e. one can train over a continuous range of scales and subsequently query at arbitrary resolutions. Due to the difficulty of performing super-resolution on inherently noisy data, we analyze network behavior under multiple denoising strategies. Lastly we compare this method to a standard convolutional decoder using both quantitative metrics and a radiologist study implemented in Voxel, our newly developed tool for web-based evaluation of medical images.