期刊
IEEE TRANSACTIONS ON MEDICAL IMAGING
卷 40, 期 11, 页码 3140-3153出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TMI.2021.3096232
关键词
Image reconstruction; Imaging; Magnetic resonance imaging; Data models; Training; Deep learning; Supervised learning; MR reconstruction; deep learning; learned data consistency; dynamic magnetic resonance imaging
类别
资金
- National Key Research and Development Program of China [2017YFC0108802, 2020YFA0712200]
- National Natural Science Foundation of China [12026603, U1805261, 61771463, 81971611, 61871373, 81729003]
- Key Field Research and Development Program of Guangdong Province [2018B030335001]
- Science and Technology Plan Program of Guangzhou [202007030002]
- Shenzhen Peacock Plan Team Program [KQTD20180413181834876]
- Innovation and Technology Commission of the Government of Hong Kong SAR [MRP/001/18X]
- Engineering Laboratory Program of Chinese Academy of Sciences [KFJ-PTXM-012]
- Strategic Priority Research Program of Chinese Academy of Sciences [XDC07040000, XDB25000000]
- China Postdoctoral Science Foundation [2020M682990, 2021M693316]
- SIAT Innovation Program for Excellent Young Researchers [E1G031]
A new DL-based approach termed Learned DC is proposed in this work, which implicitly learns data consistency for MR image reconstruction. The method is evaluated with highly undersampled dynamic data and outperforms the state-of-the-art in both quantitative and qualitative performance.
Magnetic resonance (MR) image reconstruction from undersampled k-space data can be formulated as a minimization problem involving data consistency and image prior. Existing deep learning (DL)-based methods for MR reconstruction employ deep networks to exploit the prior information and integrate the prior knowledge into the reconstruction under the explicit constraint of data consistency, without considering the real distribution of the noise. In this work, we propose a new DL-based approach termed Learned DC that implicitly learns the data consistency with deep networks, corresponding to the actual probability distribution of system noise. The data consistency term and the prior knowledge are both embedded in the weights of the networks, which provides an utterly implicit manner of learning reconstruction model. We evaluated the proposed approach with highly undersampled dynamic data, including the dynamic cardiac cine data with up to 24-fold acceleration and dynamic rectum data with the acceleration factor equal to the number of phases. Experimental results demonstrate the superior performance of the Learned DC both quantitatively and qualitatively than the state-of-the-art.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据