General Electric Company

Ge Precision Healthcare LLC

Morpho Detection, Inc.

Leland Stanford Junior University

Rensselaer Polytechnic Institute

Jed Douglas Pack

Bruno Kristiaan Bernard De Man

Zhye Yin

Peter Michael Edic

Jiang Hsieh

Kai Zeng

Xiaoye Wu

Prem Venugopal

Brian Edward Nett

Bruno Kristiaan Bernard DeMan

Samit Kumar Basu

Jiahua Fan

Soumya Ghose

Jhimli Mitra

Bernhard Erich Hermann Claus

David Allen Langan

James Vradenburg Miller

Naveen Stephan Chandra

Cynthia Elizabeth Landberg Davis

Sathish Ramani

Jie Tang

Kyle Morgan Champley

Jean-Baptiste Thibault

Chad Allan Smith

Dan Xu

Lin Fu

Bruno De Man

Maria Iatrou

Mingye Wu

Thomas Matthew Benson

Guangzhi Cao

Eri Haneda

Norbert Joseph Pelc

Ge Wang

John Eric Tkaczyk

Darin Robert Okerlund

Robert F Senzig

Abdelaziz Ikhlef

Uwe Wiedmann

Colin Richard Wilson

James Walter LeBlanc

Jeffrey Alan Kautzer

Vance Scott Robinson

Yannan Jin

Baojun Li

Xin Wang

Zhou Yu

Paavana Sainath

John Irvin Jackson

Paulo Ricardo Dos Santos Mendonca

Grant Morey Stevens

David Joseph Pitterle

Denis Perrillat-Amede

Pedro Andres Garzon

Mary Sue Kulpins

Andrea Marie Schmitz

Jean-Francois Larroux

Bruno K B De Man

Adam Budde

Todd Jason Gable

Roy Arnulf Helge Nilsen

Priti Madhav

Eugene Alex Ingerman

Xia Li

Kentaro Ogata

Ming Yan

Ken Arai

Lusik Cherkezyan

John Charles Hill

Peter Michael Edie

A computer-implemented method includes obtaining, via a processor, clinical images including vessels and generating, via the processor, straightened-out images for each coronary tree path within respective clinical images, The method also includes extracting, via the processor, segmented 3D image patches, determining, via the processor, overlapping binary mask volumes for each segment, and predicting, via the processor, pressure drops across the segmented image patches using a trained deep neural network.

System and method for deep-learning based estimation of coronary artery pressure drop

Systems and methods are provided for computed tomography (CT) imaging. In one embodiment, a method comprises adaptively blending at least two input image volumes with different spatially-variant noise characteristics to generate an output image volume with uniform noise throughout the output image volume. In this way, images may be reconstructed from projection data with data redundancy without introducing image artifacts from stitching images or variance in image noise due to the data redundancy.

Systems and methods for adaptive blending in computed tomography imaging

Methods and systems are provided for identifying motion in medical images. In one example, a method includes obtaining projection data of an imaging subject, reconstructing a first image of a location of the imaging subject from the projection data using a first reconstruction technique and reconstructing a second image corresponding to the same location of the imaging subject from the of projection data using a second reconstruction technique, different than the first reconstruction technique in terms of temporal sensitivity, calculating an inconsistency metric quantifying temporal inconsistencies between the first image and the second image, and taking an action based on the inconsistency metric.

Systems and methods for motion detection in medical images

A computer-implemented method includes obtaining, via a processor, segmented image patches of a vessel along a coronary tree path and associated coronary flow distribution for respective vessel segments in the segmented image patches. The method also includes determining, via the processor, a pressure drop distribution along an axial length of the vessel from the segmented image patches and the associated coronary flow distribution. The method further includes determining, via the processor, critical points in the pressure drop distribution. The method even further includes detecting, via the processor, a presence of a stenosis based on the critical points in the pressure drop distribution.

System and method for detecting stenosis

A computer-implemented method includes generating, via a processor, synthetic vessels. The method also includes performing, via the processor, three-dimensional (3D) computational fluid dynamics (CFD) on the synthetic vessels for different flow rates to generate 3D CFD data. The method further includes extracting, via the processor, 3D image patches from the synthetic vessels. The method even further includes obtaining, via the processor, pressure drops across the 3D image patches from the 3D CFD data. The method yet further includes training, via the processor, a deep neural network utilizing the 3D image patches, the pressure drops, and associated flow rates to generate a trained deep neural network.

Systems and techniques that facilitate hybrid unsupervised and supervised image segmentation are provided. In various embodiments, a system can access a computed tomography (CT) image depicting an anatomical structure. In various aspects, the system can generate, via an unsupervised modeling technique, at least one class probability mask of the anatomical structure based on the CT image. In various instances, the system can generate, via a deep-learning model, an image segmentation based on the CT image and based on the at least one class probability mask.

Hybrid unsupervised and supervised image segmentation model

A method includes acquiring a first dataset of projection measurements at a first energy spectrum and a second dataset of projection measurements at a second energy spectrum different from the first energy spectrum by switching between acquiring the first dataset for a set number of consecutive views at different projection angles at the first energy spectrum and acquiring the second dataset for the set number of consecutive views at different projection angles at the second energy spectrum. The set number of consecutive views is greater than one. The method includes supplementing both the first dataset with estimated projection measurements at the first energy spectrum and the second dataset with estimated projection measurements at the second energy spectrum to provide missing projection measurements at different projection angles not acquired during the imaging scan for the first dataset and the second dataset.

System and method for utilizing dual energy imaging in a computed tomography imaging system

Various methods and systems are provided for stationary CT imaging. In one embodiment, an imaging system comprises a stationary distributed x-ray source unit comprising a plurality of emitters positioned to emit x-ray beams through the imaging volume, one or more detector arrays extending around at least a portion of an imaging volume, each detector array comprising a plurality of detector elements, each detector element configured to receive x-ray beams from more than one emitter, and an anti-scatter device configured to be positioned between one or more emitters of the plurality of emitters and an object in the imaging volume.

Systems and methods for a stationary CT imaging system

Various methods and systems are provided for stationary CT imaging. In one embodiment, a method for an imaging system includes activating an emitter of a plurality of emitters of a stationary distributed x-ray source unit to emit an x-ray beam toward an object within an imaging volume, where the x-ray source unit does not rotate around the imaging volume, receiving the x-ray beam at a subset of detector elements of a plurality of detector elements of one or more detector arrays, sampling the plurality of detector elements to generate a total transmission profile, an attenuation profile, and a scatter measurement, generating a scatter-corrected attenuation profile by entering the total transmission profile, the attenuation profile, and the scatter measurement as inputs to a model, and reconstructing one or more images from the scatter-corrected attenuation profile.

The present disclosure relates to image reconstruction with favorable properties in terms of noise reduction, spatial resolution, detail preservation and computational complexity. The disclosed techniques may include some or all of: a first-pass reconstruction, a simplified datafit term, and/or a deep learning denoiser. In various implementations, the disclosed technique is portable to different CT platforms, such as by incorporating a first-pass reconstruction step.

Iterative image reconstruction framework

The present disclosure relates to training one or more neural networks for vascular vessel assessment using synthetic image data for which ground-truth data is known. In certain implementations, the synthetic image data may be based in part, or derived from, clinical image data for which ground-truth data is not known or available. Neural networks trained in this manner may be used to perform one or more of vessel segmentation, decalcification, Hounsfield unit scoring, and/or estimation of a hemodynamic parameter.

Deep learning for arterial analysis and assessment

In accordance with the present disclosure, the present technique finds a diagnostic scan timing for a non-static object (e.g., a heart or other dynamic object undergoing motion) from raw scan data, as opposed to reconstructed image data. To find the scan timing, a monitoring scan of a patient's heart is performed. In the monitoring scan, the patient dose may be limited or minimized. As the projection data is acquired during such a monitoring scan, the projection data may be subjected to sinogram analysis in a concurrent or real-time manner to determine when to start (or trigger) the diagnostic scan.

Data-based scan gating

The present approach provides a non-invasive methodology for estimation of coronary flow and/or fractional flow reserve. In certain implementations, various approaches for personalizing blood flow models of the coronary vasculature are described. The described personalization approaches involve patient-specific measurements and do not assume or rely on the resting coronary flow being proportional to myocardial mass. Consequently, there are fewer limitations in using these approaches to obtain coronary flow and/or fractional flow reserve estimates non-invasively.

Methods for personalizing blood flow models

A method for characterizing anatomical features includes receiving scanned data and image data corresponding to a subject. The scanned data comprises sinogram data. The method further includes identifying a first region in an image of the image data corresponding to a region of interest. The method also includes determining a second region in the scanned data. The second region corresponds to the first region. The method further includes identifying a sinogram trace corresponding to the region of interest. The sinogram trace comprises sinogram data present within the second region. The method includes determining a data feature of the subject based on the sinogram trace and a deep learning network. The method also includes determining a diagnostic condition corresponding to a medical condition of the subject based on the data feature.

System and method for characterizing anatomical features

Growing community of inventors

Glenville, NY, United States of America

Jed Douglas Pack