2019
DOI: 10.1109/tmi.2018.2887072
|View full text |Cite
|
Sign up to set email alerts
|

MR Image Reconstruction Using Deep Density Priors

Abstract: Algorithms for Magnetic Resonance (MR) image reconstruction from undersampled measurements exploit prior information to compensate for missing k-space data. Deep learning (DL) provides a powerful framework for extracting such information from existing image datasets, through learning, and then using it for reconstruction. Leveraging this, recent methods employed DL to learn mappings from undersampled to fully sampled images using paired datasets, including undersampled and corresponding fully sampled images, i… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1

Citation Types

0
103
0

Year Published

2019
2019
2023
2023

Publication Types

Select...
6

Relationship

0
6

Authors

Journals

citations
Cited by 147 publications
(103 citation statements)
references
References 51 publications
0
103
0
Order By: Relevance
“…The input data may not be fully used without applying the forward physical model. To address this fidelity-lacking problem, it has been proposed to treat the network output in Eq.3 as a regularization in Eq.2 using an L2 form cost to penalize the difference between the network output and the final optimized solution (Aggarwal et al, 2019;Schlemper et al, 2018;Tezcan et al, 2017):…”
Section: Theorymentioning
confidence: 99%
See 1 more Smart Citation
“…The input data may not be fully used without applying the forward physical model. To address this fidelity-lacking problem, it has been proposed to treat the network output in Eq.3 as a regularization in Eq.2 using an L2 form cost to penalize the difference between the network output and the final optimized solution (Aggarwal et al, 2019;Schlemper et al, 2018;Tezcan et al, 2017):…”
Section: Theorymentioning
confidence: 99%
“…One common approach to combine DL and the physical model of the imaging system is to use the DL model for defining an explicit regularization in the classical Bayesian MAP framework, typically via an L1 or L2 penalization (Aggarwal et al, 2019;Schlemper et al, 2018;Tezcan et al, 2017;Wang et al, 2016). However, traditional explicit regularization terms are known to offer imperfect feature descriptions and limit image quality in Bayesian reconstruction (Jin et al, 2017).…”
Section: Introductionmentioning
confidence: 99%
“…In another category of unsupervised learning approaches for fast MR imaging, they aim to learn the probability distribution of the images to be reconstructed by network training. After that, the network‐learned image priors are applied to the constrained image reconstruction framework as an explicit constraint . Recently, https://arxiv.org/find/cs/1/au:+Tezcan_K/0/1/0/all/0/1 et al trained a variational autoencoder on patches of fully sampled MR images to capture the distribution of patches and used this prior for image reconstruction .…”
Section: Introductionmentioning
confidence: 99%
“…After that, the network‐learned image priors are applied to the constrained image reconstruction framework as an explicit constraint . Recently, https://arxiv.org/find/cs/1/au:+Tezcan_K/0/1/0/all/0/1 et al trained a variational autoencoder on patches of fully sampled MR images to capture the distribution of patches and used this prior for image reconstruction . Similarly, the strategy was also adopted in the work of Zhang et al and Bigdeli et al, where they trained networks from a noisy full sampled data pair and then used them for general image restoration tasks.…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation