Deep Learning for Super-resolution Vascular Ultrasound Imaging

Sloun, Ruud Jg van; Solomon, Oren; Bruce, Matthew; Khaing, Zin Z.; Eldar, Yonina C.; Mischi, Massimo

doi:10.1109/icassp.2019.8683813

Cited by 61 publications

(39 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Convolutional neural network (CNN) was firstly used for super-resolution (SR) image reconstruction in 2014 [20], and then different networks were proposed in reconstructing image details [21], especially for natural images and face images [22]- [24]. For medical imaging, such as X-CT [25], [26], MRI [27], [28], and ultrasound imaging [29], [30], some SR methods based on deep learning have also been proposed, which improved their spatial resolution effectively.…”

Section: Introductionmentioning

confidence: 99%

Axial Super-Resolution Study for Optical Coherence Tomography Images Via Deep Learning

et al. 2020

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Axial Super-Resolution Study for Optical Coherence Tomography Images Via Deep Learning

et al. 2020

View full text Add to dashboard Cite

“…Therefore, the MB concentration is limited by the performance of the localization method used. Recent approaches (sparsity driven [19], deep learning [18]) are very promising with regard to an substantial increase in the manageable MB concentration.…”

Section: Discussionmentioning

confidence: 99%

“…− t 1 + t 2 = 0 (17) and substitutingP v using (16) gives the result = W 0 (−t 12 e −t 12 ) + t 12 (18) with t 12 = t 2 /t 1 and W 0 (x) being the main branch of Lambert's W-function [25] (e.g., available in the MATLAB software as lambertw). See the Appendix for the detailed derivation.…”

Section: Maximum Likelihood Estimatormentioning

confidence: 99%

See 1 more Smart Citation

Assessing Vessel Reconstruction in Ultrasound Localization Microscopy by Maximum Likelihood Estimation of a Zero-Inflated Poisson Model

Dencks

Piepenbrock

Schmitz

2020

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

View full text Add to dashboard Cite

In clinical applications of super-resolution ultrasound imaging, it is often not possible to achieve a full reconstruction of the microvasculature within a limited measurement time. This makes the comparison of studies and quantitative parameters of vascular morphology and perfusion difficult. Therefore, saturation models were proposed to predict adequate measurement times and estimate the degree of vessel reconstruction. Here, we derive a statistical model for the microbubble counts in super-resolution voxels by a zero-inflated Poisson (ZIP) process. In this model, voxels either belong to vessels with probability P v and count events with Poisson rate , or they are empty and remain zero. In this model, P v represents the vessel voxel density in the super-resolution image after infinite measurement time. For the parameters P v and , we give Cramér-Rao lower bounds (CRLBs) for the estimation variance and derive maximum likelihood estimators (MLEs) in a novel closedform solution. These can be calculated with knowledge of only the counts at the end of the acquisition time. The estimators are applied to preclinical and clinical data and the MLE outperforms alternative estimators proposed before. The estimated degree of reconstruction lies between 38% and 74% after less than 90 s. Vessel probability P v ranged from 4% to 20%. The rate parameter was estimated in the range of 0.5-1.3 microbubbles/voxel. For these parameter ranges, the CRLB gives standard deviations of less than 2%, which supports that the parameters can be estimated with good precision already for limited acquisition times.

show abstract

“…As such, it is about four orders of magnitude faster than the iterative FISTA scheme. Beyond speedup, Deep-ULM also yields improved performance (in particular for high concentrations), which may be attributed to its ability to learn the relation between specific interference patterns of ultrasound waves reflecting off closely spaced microbubbles and their locations (van Sloun et al 2019).…”

Section: Leveraging Deep Learningmentioning

confidence: 99%

Super-resolution Ultrasound Imaging

Christensen-Jeffries¹,

Couture²,

Dayton³

et al. 2020

Ultrasound in Medicine & Biology

Self Cite

347

173

View full text Add to dashboard Cite

The majority of exchanges of oxygen and nutrients are performed around vessels smaller than 100 mm, allowing cells to thrive everywhere in the body. Pathologies such as cancer, diabetes and arteriosclerosis can profoundly alter the microvasculature. Unfortunately, medical imaging modalities only provide indirect observation at this scale. Inspired by optical microscopy, ultrasound localization microscopy has bypassed the classic compromise between penetration and resolution in ultrasonic imaging. By localization of individual injected microbubbles and tracking of their displacement with a subwavelength resolution, vascular and velocity maps can be produced at the scale of the micrometer. Super-resolution ultrasound has also been performed through signal fluctuations with the same type of contrast agents, or through switching on and off nano-sized phase-change contrast agents. These techniques are now being applied pre-clinically and clinically for imaging of the microvasculature of the brain, kidney, skin, tumors and lymph nodes.

show abstract

Deep Learning for Super-resolution Vascular Ultrasound Imaging

Cited by 61 publications

References 10 publications

Axial Super-Resolution Study for Optical Coherence Tomography Images Via Deep Learning

Axial Super-Resolution Study for Optical Coherence Tomography Images Via Deep Learning

Assessing Vessel Reconstruction in Ultrasound Localization Microscopy by Maximum Likelihood Estimation of a Zero-Inflated Poisson Model

Super-resolution Ultrasound Imaging

Contact Info

Product

Resources

About