On Problem Formulation, Efficient Modeling and Deep Neural Networks for High-Quality Ultrasound Imaging : Invited Presentation

Perdios, Dimitris; Besson, Adrien; Martinez, Florian; Vonlanthen, Manuel; Arditi, Marcel; Thiran, Jean‐Philippe

doi:10.1109/ciss.2019.8692870

Cited by 5 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, there has been growing interest in applying deep neural networks (DNNs) to augment or replace steps of the ultrasound image formation process. For example, there is a class of deep learning approaches that improves data quality obtained from a single plane wave transmission by enhancing the beamformed data [8]- [11]. Another class of ultrasound-based deep learning approaches produces high-quality images with reduced data sampling in order to increase frame rates [12]- [19].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Nair

Washington

Tran

et al. 2020

IEEE Trans. Ultrason., Ferroelect., Freq. Contr.

View full text Add to dashboard Cite

Single plane wave transmissions are promising for automated imaging tasks requiring high ultrasound frame rates over an extended field of view. However, a single plane wave insonification typically produces suboptimal image quality. To address this limitation, we are exploring the use of deep neural networks (DNNs) as an alternative to delay-and-sum beamforming. The objectives of this work are to obtain information directly from raw channel data and to simultaneously generate both a segmentation map for automated ultrasound tasks and a corresponding ultrasound B-mode image for interpretable supervision of the automation. We focus on visualizing and segmenting anechoic targets surrounded by tissue and ignoring or de-emphasizing less important surrounding structures. DNNs trained with Field II simulations were tested with simulated, experimental phantom, and in vivo datasets that were not included during training. With unfocused input channel data (i.e., prior to the application of receive time delays), simulated, experimental phantom, and in vivo test datasets achieved mean ± standard deviation Dice similarity coefficients of 0.92 ± 0.13, 0.92±0.03, and 0.77±0.07, respectively, and generalized contrastto-noise ratios (gCNR) of 0.95±0.08, 0.93±0.08, and 0.75±0.14, respectively. With subaperture beamformed channel data and a modification to the input layer of the DNN architecture to accept these data, the fidelity of image reconstruction increased (e.g., mean gCNR of multiple acquisitions of two in vivo breast cysts ranged 0.89-0.96), but DNN display frame rates were reduced from 395 Hz to 287 Hz. Overall, the DNNs successfully translated feature representations learned from simulated data to phantom and in vivo data, which is promising for this novel approach to simultaneous ultrasound image formation and segmentation.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Nair

Washington

Tran

et al. 2020

IEEE Trans. Ultrason., Ferroelect., Freq. Contr.

View full text Add to dashboard Cite

show abstract

“…Deep learning has seen increasing popularity and success in the fields of vision and medical imaging for such tasks as classification, detection, and segmentation. However, only recently, deep learning has started being applied to physics and inverse problems [12] and the imaging pipeline [13,14]. A neural network can be viewed as a black box performing non-linear regression with a set of parameters that control the relationship of the output on the input.…”

Section: Introductionmentioning

confidence: 99%

Detecting muscle activation using ultrasound speed of sound inversion with deep learning

Feigin

Zwecker

Freedman

et al. 2020

2020 42nd Annual International Conference of the IEEE Engineering in Medicine &Amp; Biology Society (EMBC)

View full text Add to dashboard Cite

Functional muscle imaging is essential for diagnostics of a multitude of musculoskeletal afflictions such as degenerative muscle diseases, muscle injuries, muscle atrophy, and neurological related issues such as spasticity. However, there is currently no solution, imaging or otherwise, capable of providing a map of active muscles over a large field of view in dynamic scenarios.In this work, we look at the feasibility of longitudinal sound speed measurements to the task of dynamic muscle imaging of contraction or activation. We perform the assessment using a deep learning network applied to prebeamformed ultrasound channel data for sound speed inversion.Preliminary results show that dynamic muscle contraction can be detected in the calf and that this contraction can be positively assigned to the operating muscles. Potential frame rates in the hundreds to thousands of frames per second are necessary to accomplish this.

show abstract

Challenge on Ultrasound Beamforming with Deep Learning (CUBDL)

Bell

Huang

Hyun

et al. 2020

2020 IEEE International Ultrasonics Symposium (IUS)

View full text Add to dashboard Cite

On Problem Formulation, Efficient Modeling and Deep Neural Networks for High-Quality Ultrasound Imaging : Invited Presentation

Cited by 5 publications

References 12 publications

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Detecting muscle activation using ultrasound speed of sound inversion with deep learning

Challenge on Ultrasound Beamforming with Deep Learning (CUBDL)

Contact Info

Product

Resources

About