Medical Physics

New submissions

• New submissions
• Cross-lists

New submissions for Mon, 24 Jan 22

[1]  arXiv:2201.08668 [pdf, ps, other]

Title: Clipped DeepControl: deep neural network two-dimensional pulse design with an amplitude constraint layer

Authors: Mads Sloth Vinding, Torben Ellegaard Lund

Subjects: Medical Physics (physics.med-ph); Machine Learning (cs.LG); Systems and Control (eess.SY)

Advanced radio-frequency pulse design used in magnetic resonance imaging has recently been demonstrated with deep learning of (convolutional) neural networks and reinforcement learning. For two-dimensionally selective radio-frequency pulses, the (convolutional) neural network pulse prediction time (few milliseconds) was in comparison more than three orders of magnitude faster than the conventional optimal control computation. The network pulses were from the supervised training capable of compensating scan-subject dependent inhomogeneities of B0 and B+1 fields. Unfortunately, the network presented with a non-negligible percentage of pulse amplitude overshoots in the test subset, despite the optimal control pulses used in training were fully constrained. Here, we have extended the convolutional neural network with a custom-made clipping layer that completely eliminates the risk of pulse amplitude overshoots, while preserving the ability to compensate the inhomogeneous field conditions.

[2]  arXiv:2201.08794 [pdf]

Title: Hairpin RF Resonators for Transceiver Arrays with High Inter-channel Isolation and B1 Efficiency at Ultrahigh Field 7T MR Imaging

Authors: Komlan Payne, Leslie L. Ying, Xiaoliang Zhang

Comments: 17 pages, 10 figures, 2 tables

Subjects: Medical Physics (physics.med-ph); Signal Processing (eess.SP)

Sufficient electromagnetic decoupling among resonant elements in RF array coils is required to maintain the integrity of the magnetic flux density map from individual channel at ultra-high field magnetic resonance imaging. A close-fitting or high-density transceiver array for augmented performance in parallel imaging and imaging quality, leads to strong coupling between elements of the array making the decoupling a very challenging task at ultra-high fields. High impedance RF coils have demonstrated to be a prominent design method to circumvent these coupling issues. However, inherent characteristics of these coils have ramification on the B1 field efficiency and SNR or the complexity of the design. In this work, we propose a hairpin RF coil, a simple design based on high impedance technique that provides excellent decoupling performance and superior RF magnetic field efficiency compared to the current state high impedance coils. In order to validate the feasibility of the proposed hairpin RF coils, systematical studies on decoupling performance, field distribution, and SNR are performed, and the results are compared with those obtained from one high impedance RF coil namely self-decoupled RF coil. To investigate the proposed hairpin RF coil design, a 7T 8-channel head coil array using hairpin resonators is built and evaluated. The MR imaging results in a cylindrical phantom obtained from the 8-channel array demonstrated a 13 % increase in SNR field intensity of the hairpin design compared to the self-decoupled coils at 7T under the same circumstance. Furthermore, the characteristics of the hairpin RF coils are evaluated using a more realistic human head voxel model in ANSYS HFSS.

Cross-lists for Mon, 24 Jan 22

[3]  arXiv:2201.08418 (cross-list from eess.IV) [pdf, other]

Title: SoftDropConnect (SDC) -- Effective and Efficient Quantification of the Network Uncertainty in Deep MR Image Analysis

Authors: Qing Lyu, Christopher T. Whitlow, Ge Wang

Subjects: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Medical Physics (physics.med-ph)

Recently, deep learning has achieved remarkable successes in medical image analysis. Although deep neural networks generate clinically important predictions, they have inherent uncertainty. Such uncertainty is a major barrier to report these predictions with confidence. In this paper, we propose a novel yet simple Bayesian inference approach called SoftDropConnect (SDC) to quantify the network uncertainty in medical imaging tasks with gliomas segmentation and metastases classification as initial examples. Our key idea is that during training and testing SDC modulates network parameters continuously so as to allow affected information processing channels still in operation, instead of disabling them as Dropout or DropConnet does. When compared with three popular Bayesian inference methods including Bayes By Backprop, Dropout, and DropConnect, our SDC method (SDC-W after optimization) outperforms the three competing methods with a substantial margin. Quantitatively, our proposed method generates results withsubstantially improved prediction accuracy (by 10.0%, 5.4% and 3.7% respectively for segmentation in terms of dice score; by 11.7%, 3.9%, 8.7% on classification in terms of test accuracy) and greatly reduced uncertainty in terms of mutual information (by 64%, 33% and 70% on segmentation; 98%, 88%, and 88% on classification). Our approach promises to deliver better diagnostic performance and make medical AI imaging more explainable and trustworthy.

• New submissions
• Cross-lists