Abstract: Various methods and systems are provided for multi-material decomposition for computed tomography. In one embodiment, a method comprises acquiring, via an imaging system, projection data for a plurality of x-ray spectra, estimating path lengths for a plurality of materials based on the projection data and calibration data for the imaging system, iteratively refining the estimated path lengths based on a linearized model derived from the calibration data, and reconstructing material-density images for each material of the plurality of materials from the iteratively-refined estimated path lengths. By determining path-length estimates in this way without modeling the physics of the imaging system, accurate material decomposition may be performed more quickly and with less sensitivity to changes in physics of the system, and furthermore may be extended to more than two materials.

Abstract: Various methods and systems are provided for multi-material decomposition for computed tomography. In one embodiment, a method comprises acquiring, via an imaging system, projection data for a plurality of x-ray spectra, estimating path lengths for a plurality of materials based on the projection data and calibration data for the imaging system, iteratively refining the estimated path lengths based on a linearized model derived from the calibration data, and reconstructing material-density images for each material of the plurality of materials from the iteratively-refined estimated path lengths. By determining path-length estimates in this way without modeling the physics of the imaging system, accurate material decomposition may be performed more quickly and with less sensitivity to changes in physics of the system, and furthermore may be extended to more than two materials.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Abstract: A method for reconstructing an image of an object includes acquiring a set of measured projection data, reconstructing the measured projection data using a first algorithm to generate a first reconstructed image dataset, reconstructing the measured projection data using a second algorithm to generate a second reconstructed image dataset, the second algorithm being utilized to improve the temporal resolution of the second reconstructed image dataset, and generating a final image dataset using both the first and second image datasets.

Abstract: Exemplary embodiments are directed to a system and method for estimating and compensating for anode target filtration in an X-ray tube. Certain embodiments record changes in the photon flux over the life of the tube. By comparing the flux values, filtration resulting from a roughened target surface and target deposition on the tube window may be inferred. A plurality of systems and methods for comparing flux values are provided and the relative merits and complementary effects of each discussed.

Abstract: A method for reconstructing an image of an object includes acquiring a set of measured projection data, reconstructing the measured projection data using a first algorithm to generate a first reconstructed image dataset, reconstructing the measured projection data using a second algorithm to generate a second reconstructed image dataset, the second algorithm being utilized to improve the temporal resolution of the second reconstructed image dataset, and generating a final image dataset using both the first and second image datasets.

Abstract: Systems and methods are provided for acquiring and reconstructing projection data that is mathematically complete or sufficient using a computed tomography (CT) system having stationary distributed X-ray sources and detector arrays. In one embodiment, a distributed source is provided as arcuate segments offset in the X-Y plane and along the Z-axis.

Abstract: A technique is provided for the temporal interpolation of a projection data set acquired of a dynamic object, such as a heart. The projection data set is acquired using a slowly rotating gantry and a distributed X-ray source. The projection data may be interpolated at each view position to a selected instant of time, such as relative to a cardiac phase. The resulting interpolated projection data characterize the projection data at each view location at any instant in time. The set of interpolated projection data may then be reconstructed to generate images and/or volume with improved temporal resolution.

Abstract: A technique is provided for improving z-axis coverage and/or reducing cone beam artifacts during CT imaging. Multiple X-ray emission points are provided along the z-axis. Some or all of the emission points may be concurrently active. X-rays from concurrently active emission points are collimated so that X-rays from two or more emission points do not overlap on the detector. In addition, different groups of concurrently activated emission points may be sequentially or alternately activated, in conjunction with collimation, to prevent the overlap of X-rays from different emission points on the detector. In this manner, The X-rays may be timed and collimated such that the respective streams of radiation become adjacent at different locations, such as at the detector, the isocenter, or edge of the field of view.

Abstract: A system and method for ascertaining the identity of an object within an enclosed article. The system includes an acquisition subsystem utilizing a stationary radiation source and detector, a reconstruction subsystem, a computer-aided detection (CAD) subsystem, and a 2D/3D visualization subsystem. The detector may be an energy discriminating detector. The acquisition subsystem communicates view data to the reconstruction subsystem, which reconstructs it into image data and communicates it to the CAD subsystem. The CAD subsystem analyzes the image data to ascertain whether it contains any area of interest. Any such area of interest data is sent to the reconstruction subsystem for further reconstruction, using more rigorous algorithms and further analyzed by the CAD subsystem. Other information, such as risk variables or trace chemical detection information may be communicated to the CAD subsystem to be included in its analysis.

Abstract: A method for reducing artifacts in image data generated by a computed tomography system is provided. The artifacts are due to the presence of a high density object in a subject of interest. Initially, measured sinogram data is received from the computed tomography system. The sinogram data is representative of a plurality of sinogram elements. The measured sinogram data is reconstructed to generate initial reconstructed image data. A trace of the high density object is identified in the measured sinogram data. Then a region of interest is identified in the initial reconstructed image data. An optimization criterion is identified based upon the region of interest. The sinogram elements in the trace of the high density object in the measured sinogram data is iteratively adjusted based upon the optimization criterion to generate corrected sinogram data. The corrected sinogram data is reconstructed to generate improved reconstructed image data.

Abstract: A technique is provided for imaging a field of view using an X-ray source comprising two or more emission points. Each emission point is configured to emit a fan of radiation encompassing less than the entire field of view. The emission points are activated individually and rotate about the field of view, allowing respective streams of radiation to be emitted at various view angles about the field of view. The emission points, which may correspond to different radial regions of the field of view, may be differentially activated to emphasize a region of interest within the field of view. The multiple emission points may be extrapolated along the longitudinal axis in duplicate or offset configurations.

Abstract: One or more techniques are provided for adapting a reconstruction process to account for the motion of an imaged object or organ, such as the heart. In particular, projection data of the moving object or organ is acquired using a slowly rotating CT gantry. Motion data may be determined from the projection data or from images reconstructed from the projection data. The motion data may be used to reconstruct motion-corrected images from the projection data. The motion-corrected images may be associated to form motion-corrected volume renderings.

Abstract: A method and system are provided for processing an acquired image signal in parallel to generate a reconstructed image signal. In one embodiment, a processing component is provided comprising one or more field-programmable gate arrays configured as co-processors. Other aspects of the present technique provide a pipelined processor configured to forward- and back-project image data using the same data path and arithmetic units.