Abstract
The 3dimensional (3D) volume imaging in magnetic resonanceimaging (MRI) enables encoding in frequency, phaseand partition directions in k-space. To increase scan speed,subsampling of multiple-coil k-space data is applied. The imagereconstruction problem from subsampled 3D k-space datais non-convex due to the coupling between image and sensitivityprofile of coils. In this paper, it is shown that the magnitudeof 3D image is constrained in a convex hull and thus canbe solved by a two-step convex optimization procedure. Comparedwith other methods, this method directly processes 3Dk-space data without using sensitivity profile of coils. Phantomand in vivo data are used to test the proposed method,resulting in improved reconstructed image with lower meansquared error.
Original language | English |
---|---|
Title of host publication | Proceedings of the 8th International Conference on Signal Processing Systems, ICSPS 2016 |
Subtitle of host publication | Auckland, New Zealand; 21-24 November 2016 |
Place of Publication | New Yrok NY USA |
Publisher | Association for Computing Machinery (ACM) |
Pages | 41-45 |
Number of pages | 5 |
ISBN (Electronic) | 9781450347907 |
DOIs | |
Publication status | Published - 21 Nov 2016 |
Event | International Conference on Signal Processing Systems 2016 - Auckland, New Zealand Duration: 21 Nov 2016 → 24 Nov 2016 Conference number: 8th https://dl.acm.org/doi/proceedings/10.1145/3015166 |
Conference
Conference | International Conference on Signal Processing Systems 2016 |
---|---|
Abbreviated title | ICSPS 2016 |
Country/Territory | New Zealand |
City | Auckland |
Period | 21/11/16 → 24/11/16 |
Internet address |
Keywords
- Convex optimization
- Parallel magnetic resonance imaging
- Volume imaging