Title :
A sparse Bayesian learning for highly accelerated dynamic MRI
Author :
Jung, Hong ; Ye, Jong Chul
Author_Institution :
Dept. of Bio & Brain Eng., Korea Adv. Inst. of Sci. & Technol. (KAIST), Daejeon, South Korea
Abstract :
In dynamic MRI, spatio-temporal resolution is a very important issue. Recently, compressed sensing approach has become a highly attracted imaging technique since it enables accelerated acquistion without aliasing artifacts. Our group has proposed an ℓ1-norm based compressed sensing dynamic MRI called k-t FOCUSS, which outperforms existing methods. However, it is known that the restrictive conditions for ℓ1 exact reconstruction usually cost more measurements than ℓ1 minimization. In this paper, we adopts a sparse Bayesian learning approach to improve k-t FOCUSS and achieve ℓ0 solution. We demonstrated the improved image quality using in vivo cardiac cine imaging.
Keywords :
Bayes methods; biomedical MRI; image reconstruction; learning (artificial intelligence); medical image processing; sparse matrices; ℓ1 exact reconstruction; compressed sensing; highly accelerated dynamic MRI; in vivo cardiac cine imaging; k-t FOCUSS; sparse Bayesian learning; spatio-temporal resolution; Acceleration; Bayesian methods; Compressed sensing; Costs; Focusing; High-resolution imaging; Image quality; Image reconstruction; In vivo; Magnetic resonance imaging; Compressed sensing; Dynamic MRI; Sparse Bayesian learning; l1 minimization;
Conference_Titel :
Biomedical Imaging: From Nano to Macro, 2010 IEEE International Symposium on
Conference_Location :
Rotterdam
Print_ISBN :
978-1-4244-4125-9
Electronic_ISBN :
1945-7928
DOI :
10.1109/ISBI.2010.5490367