Optimal MRI Undersampling Patterns for Ultimate Benefit of Medical Vision Tasks
This work addresses the need for faster MRI scans while maintaining diagnostic accuracy for medical professionals, representing a shift from traditional compressed sensing methods.
The authors tackled the problem of accelerating MRI by optimizing undersampling patterns in k-space to improve downstream medical vision tasks like segmentation and classification, rather than focusing on image reconstruction quality. They demonstrated up to 12% improvement in Dice score for segmentation at 16x acceleration compared to other patterns.
To accelerate MRI, the field of compressed sensing is traditionally concerned with optimizing the image quality after a partial undersampling of the measurable $\textit{k}$-space. In our work, we propose to change the focus from the quality of the reconstructed image to the quality of the downstream image analysis outcome. Specifically, we propose to optimize the patterns according to how well a sought-after pathology could be detected or localized in the reconstructed images. We find the optimal undersampling patterns in $\textit{k}$-space that maximize target value functions of interest in commonplace medical vision problems (reconstruction, segmentation, and classification) and propose a new iterative gradient sampling routine universally suitable for these tasks. We validate the proposed MRI acceleration paradigm on three classical medical datasets, demonstrating a noticeable improvement of the target metrics at the high acceleration factors (for the segmentation problem at $\times$16 acceleration, we report up to 12% improvement in Dice score over the other undersampling patterns).