Fast image reconstruction with L2-regularization
Purpose We introduce L2-regularized reconstruction algorithms with closed-form solutions that achieve dramatic computational speed-up relative to state of the art L1- and L2-based iterative algorithms while maintaining similar image quality for various applications in MRI reconstruction. Materials...
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Wiley Blackwell
2015
|
| Online Access: | http://hdl.handle.net/1721.1/99708 https://orcid.org/0000-0002-4916-6314 https://orcid.org/0000-0002-7637-2914 |
| _version_ | 1826209546719199232 |
|---|---|
| author | Bilgic, Berkin Chatnuntawech, Itthi Fan, Audrey P. Setsompop, Kawin Cauley, Stephen F. Adalsteinsson, Elfar Wald, Lawrence |
| author2 | Harvard University--MIT Division of Health Sciences and Technology |
| author_facet | Harvard University--MIT Division of Health Sciences and Technology Bilgic, Berkin Chatnuntawech, Itthi Fan, Audrey P. Setsompop, Kawin Cauley, Stephen F. Adalsteinsson, Elfar Wald, Lawrence |
| author_sort | Bilgic, Berkin |
| collection | MIT |
| description | Purpose
We introduce L2-regularized reconstruction algorithms with closed-form solutions that achieve dramatic computational speed-up relative to state of the art L1- and L2-based iterative algorithms while maintaining similar image quality for various applications in MRI reconstruction.
Materials and Methods
We compare fast L2-based methods to state of the art algorithms employing iterative L1- and L2-regularization in numerical phantom and in vivo data in three applications; (i) Fast Quantitative Susceptibility Mapping (QSM), (ii) Lipid artifact suppression in Magnetic Resonance Spectroscopic Imaging (MRSI), and (iii) Diffusion Spectrum Imaging (DSI). In all cases, proposed L2-based methods are compared with the state of the art algorithms, and two to three orders of magnitude speed up is demonstrated with similar reconstruction quality.
Results
The closed-form solution developed for regularized QSM allows processing of a three-dimensional volume under 5 s, the proposed lipid suppression algorithm takes under 1 s to reconstruct single-slice MRSI data, while the PCA based DSI algorithm estimates diffusion propagators from undersampled q-space for a single slice under 30 s, all running in Matlab using a standard workstation.
Conclusion
For the applications considered herein, closed-form L2-regularization can be a faster alternative to its iterative counterpart or L1-based iterative algorithms, without compromising image quality. |
| first_indexed | 2024-09-23T14:24:11Z |
| format | Article |
| id | mit-1721.1/99708 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:24:11Z |
| publishDate | 2015 |
| publisher | Wiley Blackwell |
| record_format | dspace |
| spelling | mit-1721.1/997082022-10-01T21:05:10Z Fast image reconstruction with L2-regularization Bilgic, Berkin Chatnuntawech, Itthi Fan, Audrey P. Setsompop, Kawin Cauley, Stephen F. Adalsteinsson, Elfar Wald, Lawrence Harvard University--MIT Division of Health Sciences and Technology Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Bilgic, Berkin Chatnuntawech, Itthi Fan, Audrey P. Wald, Lawrence Adalsteinsson, Elfar Purpose We introduce L2-regularized reconstruction algorithms with closed-form solutions that achieve dramatic computational speed-up relative to state of the art L1- and L2-based iterative algorithms while maintaining similar image quality for various applications in MRI reconstruction. Materials and Methods We compare fast L2-based methods to state of the art algorithms employing iterative L1- and L2-regularization in numerical phantom and in vivo data in three applications; (i) Fast Quantitative Susceptibility Mapping (QSM), (ii) Lipid artifact suppression in Magnetic Resonance Spectroscopic Imaging (MRSI), and (iii) Diffusion Spectrum Imaging (DSI). In all cases, proposed L2-based methods are compared with the state of the art algorithms, and two to three orders of magnitude speed up is demonstrated with similar reconstruction quality. Results The closed-form solution developed for regularized QSM allows processing of a three-dimensional volume under 5 s, the proposed lipid suppression algorithm takes under 1 s to reconstruct single-slice MRSI data, while the PCA based DSI algorithm estimates diffusion propagators from undersampled q-space for a single slice under 30 s, all running in Matlab using a standard workstation. Conclusion For the applications considered herein, closed-form L2-regularization can be a faster alternative to its iterative counterpart or L1-based iterative algorithms, without compromising image quality. National Institute for Biomedical Imaging and Bioengineering (U.S.) (Grant NIBIB K99EB012107) National Institutes of Health (U.S.) (Grant NIH R01 EB007942) National Institute for Biomedical Imaging and Bioengineering (U.S.) (Grant NIBIB R01EB006847) Grant K99/R00 EB008129 National Center for Research Resources (U.S.) (Grant NCRR P41RR14075) National Institutes of Health (U.S.) (Blueprint for Neuroscience Research U01MH093765) Siemens Corporation Siemens-MIT Alliance MIT-Center for Integration of Medicine and Innovative Technology (Medical Engineering Fellowship) 2015-11-04T16:16:27Z 2015-11-04T16:16:27Z 2013-11 2013-04 Article http://purl.org/eprint/type/JournalArticle 10531807 1522-2586 http://hdl.handle.net/1721.1/99708 Bilgic, Berkin, Itthi Chatnuntawech, Audrey P. Fan, Kawin Setsompop, Stephen F. Cauley, Lawrence L. Wald, and Elfar Adalsteinsson. “Fast Image Reconstruction with L2-Regularization.” Journal of Magnetic Resonance Imaging 40, no. 1 (November 4, 2013): 181–191. https://orcid.org/0000-0002-4916-6314 https://orcid.org/0000-0002-7637-2914 en_US http://dx.doi.org/10.1002/jmri.24365 Journal of Magnetic Resonance Imaging Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Wiley Blackwell PMC |
| spellingShingle | Bilgic, Berkin Chatnuntawech, Itthi Fan, Audrey P. Setsompop, Kawin Cauley, Stephen F. Adalsteinsson, Elfar Wald, Lawrence Fast image reconstruction with L2-regularization |
| title | Fast image reconstruction with L2-regularization |
| title_full | Fast image reconstruction with L2-regularization |
| title_fullStr | Fast image reconstruction with L2-regularization |
| title_full_unstemmed | Fast image reconstruction with L2-regularization |
| title_short | Fast image reconstruction with L2-regularization |
| title_sort | fast image reconstruction with l2 regularization |
| url | http://hdl.handle.net/1721.1/99708 https://orcid.org/0000-0002-4916-6314 https://orcid.org/0000-0002-7637-2914 |
| work_keys_str_mv | AT bilgicberkin fastimagereconstructionwithl2regularization AT chatnuntawechitthi fastimagereconstructionwithl2regularization AT fanaudreyp fastimagereconstructionwithl2regularization AT setsompopkawin fastimagereconstructionwithl2regularization AT cauleystephenf fastimagereconstructionwithl2regularization AT adalsteinssonelfar fastimagereconstructionwithl2regularization AT waldlawrence fastimagereconstructionwithl2regularization |