國立虎尾科技大學 |

Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging./
作者:	Rich, Adam.
面頁冊數:	1 online resource (162 pages)
附註:	Source: Dissertation Abstracts International, Volume: 78-11(E), Section: B.
標題:	Electrical engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9781369980882

Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging.
Rich, Adam.

Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging. - 1 online resource (162 pages)

Source: Dissertation Abstracts International, Volume: 78-11(E), Section: B.

Thesis (Ph.D.)--The Ohio State University, 2017.

Includes bibliographical references

Since the digital revolution, our fascination with detail is endless. High definition images and video have become standard. Televisions and cellphones with screens containing 2-10 million pixels are commonplace. The field of magnetic resonance imaging (MRI) is no different. In MRI, doctors push for higher spatial and temporal resolution images, with current brain imaging having sub mm pixel resolution. However, acquisition time for MRI is proportional to the number of pixels in the image. Therefore, high resolution images can have unreasonably long scan times that reduce patient comfort and increase costs. Phase contrast MRI (PC-MRI) encodes velocity information into the phase of the complex valued MR image. Multiple encoded images are required to recover the velocity information, further increasing scan duration. However, due to the redundant nature of the data collection, it is possible to apply regularizing constraints to aid in the inversion of velocity maps. In this work we propose Bayesian imaging techniques to accelerate acquisition of PC-MRI data. The proposed techniques exploit structure across velocity encoded images via a mixture density prior that applies different regularization based on the presence of velocity within the image. Structure across space and time is exploited using non-decimated wavelet transforms. Using these prior-densities, we form a Bayesian data model and an associated factor graph of the posterior distribution. Through a combination of standard and loopy belief propagation, iterative inversion algorithms are derived. Computation on the loopy sections of the graph is greatly accelerated through the use of generalized approximate message passing. For 2D through-plane flow, the Bayesian techniques outperform prior art in terms of normalized mean squared error and flow quantification, e.g. stroke volume and peak velocity. In addition, the proposed technique enables image collection accelerated by a factor of 10 while preserving flow quantification. In real-time imaging, we demonstrate the feasibility of through-plane PC-MRI. For 4D flow MRI, the proposed technique enables rate 20 acceleration in a flow phantom. Finally, we demonstrate the first single breath-hold, cardiac 4D flow PC-MRI acquisition in vivo.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9781369980882Subjects--Topical Terms:

596380
Electrical engineering.
Index Terms--Genre/Form:

554714
Electronic books.

Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging.
LDR:03469ntm a2200325K 4500 001 914378
005 20180703084809.5
006 m o u
007 cr mn||||a|a||
008 190606s2017 xx obm 000 0 eng d
020 $a 9781369980882
035 $a (MiAaPQ)AAI10626881
035 $a (MiAaPQ)OhioLINK:osu1490987092511083
035 $a AAI10626881
040 $a MiAaPQ $b eng $c MiAaPQ
100 1 $a Rich, Adam. $3 1187617
245 1 0 $a Bayesian Models for Practical Flow Imaging Using Phase Contrast Magnetic Resonance Imaging.
264 0 $c 2017
300 $a 1 online resource (162 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 78-11(E), Section: B.
500 $a Adviser: Lee C. Potter.
502 $a Thesis (Ph.D.)--The Ohio State University, 2017.
504 $a Includes bibliographical references
520 $a Since the digital revolution, our fascination with detail is endless. High definition images and video have become standard. Televisions and cellphones with screens containing 2-10 million pixels are commonplace. The field of magnetic resonance imaging (MRI) is no different. In MRI, doctors push for higher spatial and temporal resolution images, with current brain imaging having sub mm pixel resolution. However, acquisition time for MRI is proportional to the number of pixels in the image. Therefore, high resolution images can have unreasonably long scan times that reduce patient comfort and increase costs. Phase contrast MRI (PC-MRI) encodes velocity information into the phase of the complex valued MR image. Multiple encoded images are required to recover the velocity information, further increasing scan duration. However, due to the redundant nature of the data collection, it is possible to apply regularizing constraints to aid in the inversion of velocity maps. In this work we propose Bayesian imaging techniques to accelerate acquisition of PC-MRI data. The proposed techniques exploit structure across velocity encoded images via a mixture density prior that applies different regularization based on the presence of velocity within the image. Structure across space and time is exploited using non-decimated wavelet transforms. Using these prior-densities, we form a Bayesian data model and an associated factor graph of the posterior distribution. Through a combination of standard and loopy belief propagation, iterative inversion algorithms are derived. Computation on the loopy sections of the graph is greatly accelerated through the use of generalized approximate message passing. For 2D through-plane flow, the Bayesian techniques outperform prior art in terms of normalized mean squared error and flow quantification, e.g. stroke volume and peak velocity. In addition, the proposed technique enables image collection accelerated by a factor of 10 while preserving flow quantification. In real-time imaging, we demonstrate the feasibility of through-plane PC-MRI. For 4D flow MRI, the proposed technique enables rate 20 acceleration in a flow phantom. Finally, we demonstrate the first single breath-hold, cardiac 4D flow PC-MRI acquisition in vivo.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Medical imaging. $3 1180167
655 7 $a Electronic books. $2 local $3 554714
690 $a 0544
690 $a 0574
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The Ohio State University. $b Electrical and Computer Engineering. $3 845530
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10626881 $z click for full text (PQDT)