Self-supervised MRI denoising: leveraging Stein's unbiased risk estimator and spatially resolved noise maps.

Thermal noise caused by the imaged object is an intrinsic limitation in magnetic resonance imaging (MRI), resulting in an impaired clinical value of the acquisitions. Recently, deep learning (DL)-based denoising methods achieved promising results by extracting complex feature representations from large data sets. Most approaches are trained in a supervised manner by directly mapping noisy to noise-free ground-truth data and, therefore, require extensive paired data sets, which can be expensive or infeasible to obtain for medical imaging applications. In this work, a DL-based denoising approach is investigated which operates on complex-valued reconstructed magnetic resonance (MR) images without noise-free target data. An extension of Stein's unbiased risk estimator (SURE) and spatially resolved noise maps quantifying the noise level with pixel accuracy were employed during the training process. Competitive denoising performance was achieved compared to supervised training with mean squared error (MSE) despite optimizing the model without noise-free target images. The proposed DL-based method can be applied for MR image enhancement without requiring noise-free target data for training. Integrating the noise maps as an additional input channel further enables the regulation of the desired level of denoising to adjust to the preference of the radiologist.

Motion guidance lines for robust data consistency-based retrospective motion correction in 2D and 3D MRI.

PURPOSE: To develop a robust retrospective motion-correction technique based on repeating k-space guidance lines for improving motion correction in Cartesian 2D and 3D brain MRI. METHODS: The motion guidance lines are inserted into the standard sequence orderings for 2D turbo spin echo and 3D MPRAGE to inform a data consistency-based motion estimation and reconstruction, which can be guided by a low-resolution scout. The extremely limited number of required guidance lines are repeated during each echo train and discarded in the final image reconstruction. Thus, integration within a standard k-space acquisition ordering ensures the expected image quality/contrast and motion sensitivity of that sequence. RESULTS: Through simulation and in vivo 2D multislice and 3D motion experiments, we demonstrate that respectively 2 or 4 optimized motion guidance lines per shot enables accurate motion estimation and correction. Clinically acceptable reconstruction times are achieved through fully separable on-the-fly motion optimizations (˜1 s/shot) using standard scanner GPU hardware. CONCLUSION: The addition of guidance lines to scout accelerated motion estimation facilitates robust retrospective motion correction that can be effectively introduced without perturbing standard clinical protocols and workflows.

Deep learning-based motion quantification from k-space for fast model-based magnetic resonance imaging motion correction.

BACKGROUND: Intra-scan rigid-body motion is a costly and ubiquitous problem in clinical magnetic resonance imaging (MRI) of the head. PURPOSE: State-of-the-art methods for retrospective motion correction in MRI are often computationally expensive or in the case of image-to-image deep learning (DL) based methods can be prone to undesired alterations of the image (hallucinations'). In this work we introduce a novel rigid-body motion correction method which combines the advantages of classical model-driven and data-consistency (DC) preserving approaches with a novel DL algorithm, to provide fast and robust retrospective motion correction. METHODS: The proposed Motion Parameter Estimating Densenet (MoPED) retrospectively estimates subject head motion during MRI acquisitions using a DL network with DenseBlocks and multitask learning. It quantifies the 2D rigid in-plane motion parameters slice-wise for each echo train (ET) of a Cartesian T2-weighted 2D Turbo-Spin-Echo sequence. The network receives a center patch of the motion corrupted k-space as well as an additional motion-free low-resolution reference scan to provide the ground truth orientation. The supervised training utilizes motion simulations based on 28 acquisitions with subject-wise training, validation, and test data splits of 70%, 23%, and 7%. During inference, MoPED is embedded in an iterative DC-driven motion correction algorithm which alternatingly updates estimates of the motion parameters and motion-corrected low-resolution k-space data. The estimated motion parameters are then used to reconstruct the final motion corrected image. The mean absolute/squared error and the Pearson correlation coefficient were used to analyze the motion parameter estimation quality on in-silico data in a quantitative evaluation. Structural similarity (SSIM), DC error and root mean squared error (RMSE) were used as metrics of image quality improvement. Furthermore, the generalization capability of the network was analyzed on two in-vivo motion volumes with 28 slices each and on one simulated T1-weighted volume. RESULTS: The motion estimation achieves a Pearson correlation of 0.968 to the simulated ground-truth of the 2433 test data slices used. In-silico results indicate that MoPED decreases the time for the optimization by a factor of around 27 compared to a conventional method and is able to reduce the RMSE of the reconstructions and average DC error by more than a factor of two compared to uncorrected images. In-vivo experiments show a decrease in computation time by a factor of around 20, a RMSE decrease from 0.055 to 0.033 and an SSIM increase from 0.795 to 0.862. Furthermore, contrast independence is demonstrated as MoPED is also able to correct T1-weighted images in simulations without retraining. Due to the model-based correction, no hallucinations were observed. CONCLUSIONS: Incorporating DL in a model-based motion correction algorithm shows great benefit on the optimization and computation time. The k-space-based estimation also allows a data consistent correction and therefore avoids the risk of hallucinations of image-to-image approaches.

Network Accelerated Motion Estimation and Reduction (NAMER): Convolutional neural network guided retrospective motion correction using a separable motion model.

PURPOSE: We introduce and validate a scalable retrospective motion correction technique for brain imaging that incorporates a machine learning component into a model-based motion minimization. METHODS: A convolutional neural network (CNN) trained to remove motion artifacts from 2D T2 -weighted rapid acquisition with refocused echoes (RARE) images is introduced into a model-based data-consistency optimization to jointly search for 2D motion parameters and the uncorrupted image. Our separable motion model allows for efficient intrashot (line-by-line) motion correction of highly corrupted shots, as opposed to previous methods which do not scale well with this refinement of the motion model. Final image generation incorporates the motion parameters within a model-based image reconstruction. The method is tested in simulations and in vivo motion experiments of in-plane motion corruption. RESULTS: While the convolutional neural network alone provides some motion mitigation (at the expense of introduced blurring), allowing it to guide the iterative joint-optimization both improves the search convergence and renders the joint-optimization separable. This enables rapid mitigation within shots in addition to between shots. For 2D in-plane motion correction experiments, the result is a significant reduction of both image space root mean square error in simulations, and a reduction of motion artifacts in the in vivo motion tests. CONCLUSION: The separability and convergence improvements afforded by the combined convolutional neural network+model-based method shows the potential for meaningful postacquisition motion mitigation in clinical MRI.