4DFlowNet: Super-Resolution 4D Flow MRI Using Deep Learning and Computational Fluid Dynamics

@inproceedings{Ferdian20204DFlowNetS4,
  title={4DFlowNet: Super-Resolution 4D Flow MRI Using Deep Learning and Computational Fluid Dynamics},
  author={Edward Ferdian and Avan Suinesiaputra and David J. Dubowitz and Debbie Zhao and Alan Q. Wang and Brett R. Cowan and Alistair A. Young},
  booktitle={Frontiers of Physics},
  year={2020}
}
4D flow magnetic resonance imaging (MRI) is an emerging imaging technique where spatiotemporal 3D blood velocity can be captured with full volumetric coverage in a single non-invasive examination. This enables qualitative and quantitative analysis of hemodynamic flow parameters of the heart and great vessels. An increase in the image resolution would provide more accuracy and allow better assessment of the blood flow, especially for patients with abnormal flows. However, this must be balanced… 

SRflow: Deep learning based super-resolution of 4D-flow MRI data

A deep convolutional neural network is proposed that learns the inter-scale relationship of the velocity vector map and leverages an efficient residual learning scheme to make it computationally feasible to super-resolve in vivo 4D-flow MRI data at a post-processing level.

Cerebrovascular super-resolution 4D Flow MRI – using deep learning to non-invasively quantify velocity, flow, and relative pressure

The approach presents a promising method to non-invasively quantify cerebrovascular hemodynamics, applicable to dedicated clinical cohorts in the future and showing potential in reducing low-resolution bias in relative pressure estimation.

Accelerated 4D‐flow MRI with 3‐point encoding enabled by machine learning

To investigate the acceleration of 4D‐flow MRI using a convolutional neural network (CNN) that produces three directional velocities from three flow encodings, without requiring a fourth reference

Super-resolution and denoising of fluid flow using physics-informed convolutional neural networks without high-resolution labels

This work presents a novel physics-informed DL-based SR solution using convolutional neural networks (CNN), which is able to produce HR flow fields from low-resolution (LR) inputs in high-dimensional parameter space by leveraging the conservation laws and boundary conditions of fluid flows.

Non-invasive hemodynamic analysis for aortic regurgitation using computational fluid dynamics and deep learning

Computational fluid dynamics simulations were transformed into synthetic 4D MRI data and used to train a variety of neural networks, which showed decreased velocity error, high structural similarity scores, and improved learning capabilities from previous work.

Synthesis of patient-specific multipoint 4D flow MRI data of turbulent aortic flow downstream of stenotic valves

We propose to synthesize patient-specific 4D flow MRI datasets of turbulent flow paired with ground truth flow data to support training of inference methods. Turbulent blood flow is computed based on

Physics-Driven Deep Learning for Computational Magnetic Resonance Imaging

An overview of the recent developments in incorporating physics information into learning-based MRI reconstruction, covering physics-driven loss functions, plug-and-play methods, generative models, and unrolled networks is provided.

Deep Learning for Computational Hemodynamics: A Brief Review of Recent Advances

It is suggested that incorporating physiologic understandings and underlying fluid mechanics laws in deep learning models will soon lead to a paradigm shift in the development novel non-invasive computational medical decisions.

Flow field tomography with uncertainty quantification using a Bayesian physics-informed neural network

We report a new approach to flow field tomography that uses the Navier–Stokes and advection–diffusion equations to regularize reconstructions. Tomography is increasingly employed to infer 2D or 3D

References

SHOWING 1-10 OF 42 REFERENCES

New imaging tools in cardiovascular medicine: computational fluid dynamics and 4D flow MRI

Blood flow imaging can provide useful information for developing predictive medicine in cardiovascular diseases, and may lead to breakthroughs in cardiovascular surgery, especially in the decision-making process.

Computational fluid dynamics simulations of blood flow regularized by 3D phase contrast MRI

This study develops a proof-of-concept numerical procedure for constructing a simulated flow field that is influenced by both direct PC-MRI measurements and a fluid physics model, thereby taking advantage of both the accuracy of PC- MRI and the high spatial resolution of CFD.

4D flow cardiovascular magnetic resonance consensus statement

It is described that 4D Flow CMR can be clinically advantageous because placement of a single acquisition volume is straightforward and enables flow through any plane across it to be calculated retrospectively and with good accuracy.

Brain MRI super resolution using 3D deep densely connected neural networks

A new neural network architecture, 3D Densely Connected Super-Resolution Networks (DCSRN) is introduced to restore HR features of structural brain MR images and it is demonstrated that this network outperforms bicubic interpolation as well as other deep learning methods in restoring 4× resolution-reduced images.

Improving visualization of 4D flow cardiovascular magnetic resonance with four-dimensional angiographic data: generation of a 4D phase-contrast magnetic resonance CardioAngiography (4D PC-MRCA)

This study proposes a novel technique, called Phase-Contrast Magnetic Resonance CardioAngiography (4D PC-MRCA), that utilizes the full potential of 4D Flow CMR when generating temporally resolvedPC-MRA data to improve visualization of the heart and major vessels throughout the cardiac cycle.

Congenital heart disease assessment with 4D flow MRI

An overview of MRI evaluation of congenital heart disease by means of example of three relatively common representative conditions: tetralogy of Fallot, aortic coarctation, and anomalous pulmonary venous drainage is provided.

Brain MRI super-resolution using 3D generative adversarial networks

An adversarial learning approach to generate high resolution MRI scans from low resolution images based on the SRGAN model that adopts 3D convolutions to exploit volumetric information is proposed.

Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network

This paper presents the first convolutional neural network capable of real-time SR of 1080p videos on a single K2 GPU and introduces an efficient sub-pixel convolution layer which learns an array of upscaling filters to upscale the final LR feature maps into the HR output.