Learning Sub-Sampling and Signal Recovery With Applications in Ultrasound Imaging

Huijben, Iris A. M.; Veeling, Bastiaan S.; Janse, Kees; Mischi, Massimo; Sloun, Ruud J. G. van

doi:10.1109/tmi.2020.3008501

Cited by 50 publications

(34 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In medical ultrasound, several strategies have been proposed to restore high image quality while reducing data sampling, transmission, and processing [28], [29], [31], [43]. With the exception of a single preliminary study reporting deep learning of color Doppler images [44], however, CNNs have not been applied as extensively to ultrasound imaging of blood flows. We proposed a deep learning platform for the direct reconstruction of power Doppler images from a 3-D space of sparse compound ultrasound data (Fig.…”

Section: Discussionmentioning

confidence: 99%

Deep-fUS: A Deep Learning Platform for Functional Ultrasound Imaging of the Brain Using Sparse Data

Ianni

Airan

2022

IEEE Trans. Med. Imaging

View full text Add to dashboard Cite

Functional ultrasound (fUS) is a rapidly emerging modality that enables whole-brain imaging of neural activity in awake and mobile rodents. To achieve sufficient blood flow sensitivity in the brain microvasculature, fUS relies on long ultrasound data acquisitions at high frame rates, posing high demands on the sampling and processing hardware. Here we develop an image reconstruction method based on deep learning that significantly reduces the amount of data necessary while retaining imaging performance. We trained convolutional neural networks to learn the power Doppler reconstruction function from sparse sequences of ultrasound data with compression factors of up to 95%. High-quality images from in vivo acquisitions in rats were used for training and performance evaluation. We demonstrate that time series of power Doppler images can be reconstructed with sufficient accuracy to detect the small changes in cerebral blood volume (~10%) characteristic of task-evoked cortical activation, even though the network was not formally trained to reconstruct such image series. The proposed platform may facilitate the development of this neuroimaging modality in any setting where dedicated hardware is not available or in clinical scanners.

show abstract

Section: Discussionmentioning

confidence: 99%

Deep-fUS: A Deep Learning Platform for Functional Ultrasound Imaging of the Brain Using Sparse Data

Ianni

Airan

2022

IEEE Trans. Med. Imaging

View full text Add to dashboard Cite

show abstract

“…In the last few years, deep CNNs have been the object of increasing attention in anatomical ultrasound image reconstruction, with a particular emphasis on models that aim to restore high-end image quality while reducing data sampling, transmission, and processing 25,26, 28,37 . With the exception of a single study using deep learning of color Doppler images 38 , however, CNNs have not been applied as extensively to blood flow imaging, and to the best of our knowledge this is the first attempt to implement an end-to-end network to perform power Doppler reconstruction from sparse ultrasonic datasets (Fig. 1).…”

Section: Discussionmentioning

confidence: 99%

“…In medical ultrasound, several strategies have been proposed to restore high image quality while reducing data sampling, transmission, and processing 30,31,33,38 . With the exception of a single preliminary study reporting deep learning of color Doppler images 39 , however, CNNs have not been applied as extensively to ultrasound imaging of blood flows. We demonstrated a deep learning method for the direct reconstruction of power Doppler images from a 3-D space of sparse compound ultrasound data (Fig.…”

Section: Discussionmentioning

confidence: 99%

Deep-fUS: Functional ultrasound imaging of the brain using deep learning and sparse data

Ianni

Airan

2020

Preprint

View full text Add to dashboard Cite

Functional ultrasound imaging is rapidly establishing itself as a state-of-the-art neuroimaging modality owing to its ability to image neural activation in awake and mobile small animals, its relatively low cost, and its unequaled portability. To achieve high blood flow sensitivity in the brain microvasculature, functional ultrasound relies on long sequences of ultrasound data acquired at high frame rates, which pose high demands on the hardware architecture and effectively limit the practical utility and clinical translation of this imaging modality. Here we propose Deep-fUS, a deep learning approach that aims to significantly reduce the amount of ultrasound data necessary, while retaining the imaging performance. We trained a neural network to learn the power Doppler reconstruction function from sparse sequences of compound ultrasound data, using ground truth images from high-quality in vivo acquisitions in rats, and with a custom loss function. The network produces highly accurate images with restored sensitivity in the smaller blood vessels even when using heavily under-sampled data. We tested the network performance in a task-evoked functional neuroimaging application, demonstrating that time series of power Doppler images can be reconstructed with adequate accuracy to compute functional activity maps. Notably, the network reduces the occurrence of motion artifacts in awake functional ultrasound imaging experiments. The proposed platform can facilitate the development of this neuroimaging modality in any setting where dedicated hardware is not available or even in clinical scanners, opening the way to new potential applications based on this technology. To our knowledge, this is the first attempt to implement a convolutional neural network approach for power Doppler reconstruction from sparse ultrasonic datasets.

show abstract

“…Modeling the sensing matrix accurately is often difficult due to the nature of US scattering 18 . Moreover, an adequate sparsifying basis has to be formalized for these CS‐based approaches, which is often difficult because of the characteristic speckles contained in US images 18,19 . Beamforming techniques have been developed to reconstruct 3D US images from a limited number of measurements, where a single focused line is virtually modeled using multiple adjacent transducer elements 20–23 …”

Section: Introductionmentioning

confidence: 99%

“…Studies have demonstrated deep learning‐based approaches significantly outperform over CS‐based methods for image reconstruction 39‐42 . Deep learning‐based US image reconstruction algorithms and US beamforming methods have been proposed for processing both fully sampled and subsampled US RF data 18,19,43 . While it is encouraging, current deep learning‐based 3D US image reconstruction methods using limited number of measurements require large amount of paired data (raw data with ground truth images) for training a usable neural network, as well as accessibility to US RF data.…”

Section: Introductionmentioning

confidence: 99%

Self‐supervised learning for accelerated 3D high‐resolution ultrasound imaging

et al. 2021

View full text Add to dashboard Cite

Purpose Ultrasound (US) imaging has been widely used in diagnosis, image‐guided intervention, and therapy, where high‐quality three‐dimensional (3D) images are highly desired from sparsely acquired two‐dimensional (2D) images. This study aims to develop a deep learning‐based algorithm to reconstruct high‐resolution (HR) 3D US images only reliant on the acquired sparsely distributed 2D images. Methods We propose a self‐supervised learning framework using cycle‐consistent generative adversarial network (cycleGAN), where two independent cycleGAN models are trained with paired original US images and two sets of low‐resolution (LR) US images, respectively. The two sets of LR US images are obtained through down‐sampling the original US images along the two axes, respectively. In US imaging, in‐plane spatial resolution is generally much higher than through‐plane resolution. By learning the mapping from down‐sampled in‐plane LR images to original HR US images, cycleGAN can generate through‐plane HR images from original sparely distributed 2D images. Finally, HR 3D US images are reconstructed by combining the generated 2D images from the two cycleGAN models. Results The proposed method was assessed on two different datasets. One is automatic breast ultrasound (ABUS) images from 70 breast cancer patients, the other is collected from 45 prostate cancer patients. By applying a spatial resolution enhancement factor of 3 to the breast cases, our proposed method achieved the mean absolute error (MAE) value of 0.90 ± 0.15, the peak signal‐to‐noise ratio (PSNR) value of 37.88 ± 0.88 dB, and the visual information fidelity (VIF) value of 0.69 ± 0.01, which significantly outperforms bicubic interpolation. Similar performances have been achieved using the enhancement factor of 5 in these breast cases and using the enhancement factors of 5 and 10 in the prostate cases. Conclusions We have proposed and investigated a new deep learning‐based algorithm for reconstructing HR 3D US images from sparely acquired 2D images. Significant improvement on through‐plane resolution has been achieved by only using the acquired 2D images without any external atlas images. Its self‐supervision capability could accelerate HR US imaging.

show abstract

Learning Sub-Sampling and Signal Recovery With Applications in Ultrasound Imaging

Cited by 50 publications

References 38 publications

Deep-fUS: A Deep Learning Platform for Functional Ultrasound Imaging of the Brain Using Sparse Data

Deep-fUS: A Deep Learning Platform for Functional Ultrasound Imaging of the Brain Using Sparse Data

Deep-fUS: Functional ultrasound imaging of the brain using deep learning and sparse data

Self‐supervised learning for accelerated 3D high‐resolution ultrasound imaging

Contact Info

Product

Resources

About