Deep Generalization of Structured Low-Rank Algorithms (Deep-SLR)

TL;DR

This paper introduces a deep learning-based method to accelerate structured low-rank algorithms for MRI reconstruction, achieving similar accuracy with significantly reduced computational time and increased robustness to motion errors.

Contribution

The paper presents a CNN-based approach that pre-learns annihilation relations, greatly reducing computation compared to traditional SLR methods and enabling calibration-less, motion-insensitive MRI reconstruction.

Findings

Achieves around 1000x faster MRI reconstruction

Performs comparably to traditional SLR schemes in accuracy

Offers robustness to motion errors and higher acceleration capabilities

Abstract

Structured low-rank (SLR) algorithms, which exploit annihilation relations between the Fourier samples of a signal resulting from different properties, is a powerful image reconstruction framework in several applications. This scheme relies on low-rank matrix completion to estimate the annihilation relations from the measurements. The main challenge with this strategy is the high computational complexity of matrix completion. We introduce a deep learning (DL) approach to significantly reduce the computational complexity. Specifically, we use a convolutional neural network (CNN)-based filterbank that is trained to estimate the annihilation relations from imperfect (under-sampled and noisy) k-space measurements of Magnetic Resonance Imaging (MRI). The main reason for the computational efficiency is the pre-learning of the parameters of the non-linear CNN from exemplar data, compared to SLR schemes that learn the linear filterbank parameters from the dataset itself. Experimental comparisons show that the proposed scheme can enable calibration-less parallel MRI; it can offer performance similar to SLR schemes while reducing the runtime by around three orders of magnitude. Unlike pre-calibrated and self-calibrated approaches, the proposed uncalibrated approach is insensitive to motion errors and affords higher acceleration. The proposed scheme also incorporates image domain priors that are complementary, thus significantly improving the performance over that of SLR schemes.