Carestream Health, Inc.

Levon O Vogelsang

Richard Alan Simon

Lawrence Allen Ray

William J Sehnert

Xiaohui Wang

David H Foos

Yuan Lin

Nathan J Packard

John Yorkston

Lori L Barski

Robert Allen Senn

Karin Toepfer

A machine learning network is trained to generate a noise field image from a radiographic (x-ray) image. The training includes accessing a number of previously acquired radiographic images, duplicating the previously acquired radiographic images, and conditioning each of the duplicated images with simulated noise content to form a plurality of simulated low-exposure images. Each of the simulated low-exposure images is paired with its corresponding previously acquired image to form a learning pair. The machine learning network is trained to generate a noise field image using the learning pairs of images. A noise suppressed image of an object can be generated by applying a scaling factor to at least a portion of the corresponding noise field image and combining the scaled noise field image with a current captured image of the object.

Noise suppression using deep convolutional networks

A method for radiographic imaging of a patient directs imaging radiation to an anatomy of the patient using a radiographic imaging apparatus. A first image of the patient anatomy is captured in a digital detector, and one or more image acquisition factors are electronically stored. A second radiographic image of the same patient anatomy is captured by accessing and using the stored one or more image acquisition factors. A combination of the first and second radiographic images is displayed or transmitted.

Physiological analysis from video x-ray imaging

A mobile radiography apparatus is configured to sparsely sample radiographic projection images to generate high resolution tomosynthesis volume images using a digital radiographic detector that is mechanically uncoupled from the x-ray source and an artificial intelligence network. The artificial intelligence network is trained to correct a volume image generated from sparsely sample projection images to generate the high resolution tomosynthesis volume images.

Low dose digital tomosynthesis system and method using artificial intelligence

A mobile radiography apparatus includes radiopaque markers disposed in a radiation path that extends from an x-ray source to a digital radiographic detector, which detector is mechanically uncoupled from the x-ray source or x-ray sources. A processing system calculates a position of the detector relative to the x-ray source or x-ray sources according to identified marker positions in acquired x-ray projection images, and reconstructs a volume image according to the acquired x-ray projection images.

Self-calibrating technique for x-ray imaging scanners

An imaging method, accesses a set of low-energy projection images and performs a low-energy reconstruction using the low-energy projection images. A synthesized intermediate low-energy projection image is generated. A high-energy reconstruction is performed using a set of high-energy projection images. A synthesized intermediate high-energy projection image is generated. A dual-energy reconstruction is performed using at least one low-energy projection image, the synthesized intermediate low-energy projection image, at least one high-energy projection image, and the synthesized intermediate high-energy projection image.

Imaging system and method for dual-energy and computed tomography

A computer implemented method for reconstructing a 3-D volume image using a radiographic imaging system having one or more x-ray sources and a digital detector. A plurality of radiographic images of a subject at various angles are captured in the digital detector. Image data in two or more pixels of the detector that are adjacent to each other in a row direction or a column direction are combined, while pixels adjacent in the other direction are not combined.

Tomographic image acquisition using asymmetrical pixel binning

A method for geometric calibration of a radiography apparatus disposes at least one radio-opaque marker in the field of view of the radiography apparatus. A series of tomosynthesis projection images of patient anatomy is acquired from the detector with the x-ray source at different positions along a scan path. For at least three projection images showing the position of the radio-opaque marker, the spatial and angular geometry of the x-ray source and detector are calculated according to the positions of the marker. A tomosynthesis image is reconstructed according to the calculated geometry. A rendering of the reconstructed image is displayed.

Fiducial marker for geometric calibration of bed-side mobile tomosynthesis system

A method for geometric calibration of a radiography apparatus acquires tomosynthesis projection images of patient anatomy from a detector and an x-ray source translated along a scan path. An epipolar geometry is calculated according to the relative position of the detector to the scan path by estimating a direction for epipolar lines that extend along an image plane that includes the detector. A region of interest of the patient anatomy that is a portion of each projection image and is fully included within each projection image in the series is defined. A consistency metric is calculated for estimated epipolar lines extending within the ROI. The method iteratively adjusts the epipolar line estimation until the consistency metric indicates accuracy to within a predetermined threshold. A portion of a tomosynthesis volume is reconstructed and displayed.

Mobile radiography calibration for tomosynthesis using epipolar data consistency

A method for geometric calibration of a mobile radiography apparatus, executed at least in part by a computer, acquires a series of tomosynthesis projection images of a patient positioned between an x-ray source of the mobile radiography apparatus and a detector that is positionally uncoupled from the x-ray source. A vector field is generated having a first set of vectors indicative of feature movement between a first acquired projection image and a second acquired projection image. The generated vector field is associated with an epipolar geometry according to an optimization of an energy relationship between an epipolar model and the generated vector field values. The mobile radiography apparatus is calibrated according to the associated model epipolar geometry. At least a portion of the tomosynthesis image is reconstructed and displayed.

Mobile radiography calibration for tomosynthesis using epipolar geometry

A method for tomosynthesis volume reconstruction acquires at least a prior projection image of the subject at a first angle and a subsequent projection image of the subject at a second angle. A synthetic image corresponding to an intermediate angle between the first and second angle is generated by a repeated process of relating an area of the synthetic image to a prior patch on the prior projection image and to a subsequent patch on the subsequent projection image according to a bidirectional spatial similarity metric, wherein the prior patch and subsequent patch have n&#xd7;m pixels; and combining image data from the prior patch and the subsequent patch to form a portion of the synthetic image. The generated synthetic image is displayed, stored, processed, or transmitted.

Virtual projection images for tomosynthesis artifact reduction

A method for imaging a subject obtains a first set of acquired projection images of a subject volume, wherein each projection image in the first set has a corresponding acquisition angle and forms an initial reconstructed volume image. A second set of synthetic projection images is generated according to processing of the acquired projection images and combined with the first set to form a combined set of projection images. The method augments the initial reconstructed image to form an improved reconstructed image by at least a first iteration of an iterative reconstruction process using the initial reconstructed image with the combined set of acquired and synthetic projection images and at least a subsequent iteration of the iterative reconstruction process using the first set of acquired projection images and fewer than, or none of, the second set of synthetic projection images. The improved reconstruction image is rendered on a display.

Iterative volume image reconstruction using synthetic projection images

An imaging method, accesses cone beam computed tomography (CBCT) data and displays, on a display monitor, at least one view of the CBCT data. The method provides an interface for a user to indicate a tomosynthesis reconstruction plane on the displayed view of the CBCT data; and displays a tomosynthesis image on the display monitor according to the indicated tomosynthesis reconstruction plane.

Tomosynthesis views from cone beam computed tomography data

A method for forming an image reconstructs a volume image according to X-ray projection images acquired at acquisition angles. The full volume image is partitioned to form at least a first and a second non-overlapping sub-volume. Within each sub-volume, forward projection images for the sub-volume are calculated, with the corresponding forward projection images computed at the acquisition angles, and with intermediate forward projection images at angles between the acquisition angles. A weight factor relates the contribution of each pixel in the X-ray projection images to each sub-volume at each acquisition angle. Synthesized sub-volume projection images are formed according to the calculated weight factors and acquired projection images in each sub-volume. Synthesized sub-volume projection images form synthesized projection images for the full volume image. A second volume image is reconstructed according to the acquired X-ray projection images and the synthesized projection images. The reconstructed second volume image is displayed, stored, or transmitted.

Volume image reconstruction using projection decomposition

A method of generating a synthesized digital projection image, executed at least in part by a computer, acquires x-ray projection images at corresponding acquisition angles, and reconstructs a first volume image. Forward projection generates a preceding forward projection image at a first acquisition angle; a following forward projection image at an adjacent second acquisition angle; and an intermediate forward projection image for an intermediate angle. A synthesized projection image is formed by identifying a candidate patch of pixels from the intermediate forward projection image, matching a first patch of pixels from the preceding forward projection image and identifying a corresponding first acquired patch of pixels from the first acquired x-ray projection image, matching a second patch of pixels from the following forward projection image and matching a second acquired patch of pixels from the forward projection image. The first and second acquired patches are combined.

Growing community of inventors

Webster, NY, United States of America

Levon O Vogelsang