Abstract
Inspired by recent advances in deep learning, we propose a framework for reconstructing dynamic sequences of 2-D cardiac magnetic resonance (MR) images from undersampled data using a deep cascade of convolutional neural networks (CNNs) to accelerate the data acquisition process. In particular, we address the case where data are acquired using aggressive Cartesian undersampling. First, we show that when each 2-D image frame is reconstructed independently, the proposed method outperforms state-of-the-art 2-D compressed sensing approaches, such as dictionary learning-based MR image reconstruction, in terms of reconstruction error and reconstruction speed. Second, when reconstructing the frames of the sequences jointly, we demonstrate that CNNs can learn spatio-temporal correlations efficiently by combining convolution and data sharing approaches. We show that the proposed method consistently outperforms state-of-the-art methods and is capable of preserving anatomical structure more faithfully up to 11-fold undersampling. Moreover, reconstruction is very fast: each complete dynamic sequence can be reconstructed in less than 10 s and, for the 2-D case, each image frame can be reconstructed in 23 ms, enabling real-time applications.
Highlights
I N MANY clinical scenarios, medical imaging is an indispensable diagnostic and research tool
One of the contributions of our work is to explore the application of convolutional neural networks (CNNs) in undersampled magnetic resonance (MR) reconstruction and investigate whether they can exploit data redundancy through learned representations
We propose a deep network architecture which forms a cascade of CNNs
Summary
I N MANY clinical scenarios, medical imaging is an indispensable diagnostic and research tool. An alternative approach is to exploit sources of explicit redundancy of the data to turn the initially underdetermined problem arising from undersampling into a determined or overdetermined problem that is solved This is the fundamental assumption underlying parallel imaging [7]. One of the contributions of our work is to explore the application of CNNs in undersampled MR reconstruction and investigate whether they can exploit data redundancy through learned representations. Reconstruction of undersampled MR images is challenging because the images typically have low signal-to-noise ratio, yet often high-quality reconstructions are needed for clinical applications To resolve this issue, we propose a deep network architecture which forms a cascade of CNNs.. Sequences can be reconstructed within 10s, which is reasonably fast for off-line reconstruction methods
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
