# Linear Predictability in MRI Reconstruction: Leveraging Shift-Invariant Fourier Structure for Faster and Better Imaging

arXiv: https://arxiv.org/abs/1903.03141 · HTML: https://mlsanity.com/paper/1903.03141/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Mar 2019.

## Built on by (most cited first, 3 of 3)

- [Optimizing Sampling Patterns for Compressed Sensing MRI with Diffusion Generative Models](https://mlsanity.com/paper/2306.03284/) (2023)
- [Physics-Informed DeepMRI: Bridging the Gap from Heat Diffusion to k-Space Interpolation](https://mlsanity.com/paper/2308.15918/) (2023)
- [A New k-Space Model for Non-Cartesian Fourier Imaging](https://mlsanity.com/paper/2505.05647/) (2025)

## Abstract (arXiv)

Over the past several decades, many different types of computational imaging approaches have been proposed for improving MRI. In this paper, we provide an overview of methods that assume that MRI Fourier data is linearly predictable. Linear prediction is well known in signal processing and may be most recognizable for its usefulness in speech processing and spectrum estimation applications. In MRI, linear predictability implies that data can be sampled below the conventional Nyquist rate, since unmeasured data may be imputed as a shift-invariant linear combination of measured samples. Linear predictive methods include some of the earliest methods in the computational MRI reconstruction field, some of the most widely utilized computational MRI methods in modern clinical practice, and some of the most flexible and versatile modern computational imaging approaches that are enabling unprecedented new styles of data acquisition. In addition, the concept of linear predictability can be used to unify a number of more classical MRI reconstruction constraints, but without needing to make the strong assumptions of classical constrained reconstruction methods.