Abstract: One or more systems, devices, computer program products and/or computer-implemented methods of use provided herein relate to image enhancement using a generative adversarial network (GAN). The computer-implemented system can comprise a memory that can store computer-executable components. The computer-implemented system can further comprise a processor that can execute the computer-executable components stored in the memory, wherein the computer-executable components can comprise a training component that can train a discriminator of the GAN to score a texture of a CT image, wherein the texture can be derived from a difference of two conditionally independent estimates produced by a generator of the GAN by respectively processing two independent noisy samples of images.

Abstract: Systems and methods for computed tomography imaging are provided. In one embodiment, a method includes acquiring an image, inputting the image to a machine learning model to generate a denoised image, the machine learning model trained with a loss function that weights variance differently from bias, and outputting the denoised image. In this way, structural details in denoised CT images may be improved while maintaining textural information in the denoised images.

Abstract: The present approach relates to the training of a machine learning algorithm for image generation and use of such a trained algorithm for image generation. Training the machine learning algorithm may involve using multiple images produced from a single set of tomographic projection or image data (such as a simple reconstruction and a computationally intensive reconstruction), where one image is the target image that exhibits the desired characteristics for the final result. The trained machine learning algorithm may be used to generate a final image corresponding to a computationally intensive algorithm from an input image generated using a less computationally intensive algorithm.

Abstract: A tomography system having a central processing unit, a system memory communicatively connected to the central processing unit, and a hardware acceleration unit communicatively connected to the central processing unit and the system memory, the hardware accelerator configured to perform at least a portion of an MBIR process on computer tomography data. The hardware accelerator unit may include one or more voxel evaluation modules which evaluate an updated value of a voxel given a voxel location in a reconstructed volume. By processing voxel data for voxels in a voxel neighborhood, processing time is reduces.

Abstract: Light reflected from an illuminated object is mixed with a reference beam and sensed at a sensor array of a digital hologram apparatus. Digital hologram data, determined from the sensed light, is dependent upon complex valued reflection coefficients of the object and upon phase perturbations in propagation paths between the object and the sensor array. Reflectance values, which may be dependent upon expected values of the absolute square of the reflection coefficients, and phase perturbations are determined for which a test function is at an extremum, where the test function contains a data fidelity term dependent upon the hologram data from a single hologram, a first regularization term dependent upon the phase perturbations and a second regularization term dependent upon the reflectance values. An image of the object may be formed from the reflectance values and a wavefront of the reflected light may be determined from the phase perturbations.

Abstract: Light reflected from an illuminated object is mixed with a reference beam and sensed at a sensor array of a digital hologram apparatus. Digital hologram data, determined from the sensed light, is dependent upon complex valued reflection coefficients of the object and upon phase perturbations in propagation paths between the object and the sensor array. Reflectance values, which may be dependent upon expected values of the absolute square of the reflection coefficients, and phase perturbations are determined for which a test function is at an extremum, where the test function contains a data fidelity term dependent upon the hologram data from a single hologram, a first regularization term dependent upon the phase perturbations and a second regularization term dependent upon the reflectance values. An image of the object may be formed from the reflectance values and a wavefront of the reflected light may be determined from the phase perturbations.

Abstract: A tomography system having a central processing unit, a system memory communicatively connected to the central processing unit, and a hardware acceleration unit communicatively connected to the central processing unit and the system memory, the hardware accelerator configured to perform at least a portion of an MBIR process on computer tomography data. The hardware accelerator unit may include one or more voxel evaluation modules which evaluate an updated value of a voxel given a voxel location in a reconstructed volume. By processing voxel data for voxels in a voxel neighborhood, processing time is reduces.

Abstract: An iterative reconstruction approach is provided that allows the use of differing weights in pixels or larger sub-regions in the reconstructed image. By way of example, the relative significance of each projection measurement may be determined based on both the measurement position and the location of the reconstructed pixel. Computationally, the significance of each projection based on these two factors is represented by a weight factor employed in the algorithmic computation.

Abstract: The present approach relates to the training of a machine learning algorithm for image generation and use of such a trained algorithm for image generation. Training the machine learning algorithm may involve using multiple images produced from a single set of tomographic projection or image data (such as a simple reconstruction and a computationally intensive reconstruction), where one image is the target image that exhibits the desired characteristics for the final result. The trained machine learning algorithm may be used to generate a final image corresponding to a computationally intensive algorithm from an input image generated using a less computationally intensive algorithm.

Abstract: A method of creating a hierarchical dictionary comprises, with a processor, extracting a number of symbols from a first image, constructing a number of refinement dictionary entries based on the symbols, the refinement dictionary entries forming a refinement dictionary, grouping a number of the refinement dictionary entries into clusters to form a number of refinement dictionary entry clusters, and constructing a number of direct dictionary entries for each of the refinement dictionary entry clusters, the direct dictionary entries forming a direct dictionary.

Abstract: A system and method for text line detection are described Examples include detection of symbols in an image received from an image-capturing device. In examples, for each of at least some of the symbols, neighboring symbols within a local region a given distance from the symbol are analyzed in order to determine a direction for a line in the local region. In examples, based on the determined directions for the lines, text lines in the image are identified.

Abstract: A tomography system having a central processing unit, a system memory communicatively connected to the central processing unit, and a hardware acceleration unit communicatively connected to the central processing unit and the system memory, the hardware accelerator configured to perform at least a portion of an MBIR process on computer tomography data. The hardware accelerator unit may include one or more voxel evaluation modules which evaluate an updated value of a voxel given a voxel location in a reconstructed volume. By processing voxel data for voxels in a voxel neighborhood, processing time is reduces.

Abstract: A system and method for text line detection are described Examples include detection of symbols in an image received from an image-capturing device. In examples, for each of at least some of the symbols, neighboring symbols within a local region a given distance from the symbol are analyzed in order to determine a direction for a line in the local regions In examples, based on the determined directions for the lines, text lines in the image are identified.

Abstract: A method includes obtaining spectral computed tomography (CT) information via an acquisition unit having an X-ray source and a CT detector. The method also includes, generating, with one or more processing units, using at least one image transform, a first basis image and a second basis image using the spectral CT information. Further, the method includes performing, with the one or more processing units, guided processing on the second basis image using the first basis image as a guide to provide a modified second basis image. Also, the method includes performing at least one inverse image transform using the first basis image and the modified second basis image to generate at least one modified image.

Abstract: A framework for an iterative reconstruction algorithm is described which combines two or more of an ordered subset method, a preconditioner method, and a nested loop method. In one type of implementation a nested loop (NL) structure is employed where the inner loop sub-problems are solved using ordered subset (OS) methods. The inner loop may be solved using OS and a preconditioner method. In other implementations, the inner loop problems are created by augmented Lagrangian methods and then solved using OS method.

Abstract: Methods and systems for model-based image processing are provided. One method includes selecting at least one reference image from a plurality of reference images, partitioning the at least one reference image into a plurality of patches, generating a probability distribution for each of the patches, and generating a model of a probability distribution for the at least one reference image using the probability distributions for each of the patches.

Abstract: A method includes obtaining spectral computed tomography (CT) information via an acquisition unit having an X-ray source and a CT detector. The method also includes, generating, with one or more processing units, using at least one image transform, a first basis image and a second basis image using the spectral CT information. Further, the method includes performing, with the one or more processing units, guided processing on the second basis image using the first basis image as a guide to provide a modified second basis image. Also, the method includes performing at least one inverse image transform using the first basis image and the modified second basis image to generate at least one modified image.

Abstract: A tomography system having a central processing unit, a system memory communicatively connected to the central processing unit, and a hardware acceleration unit communicatively connected to the central processing unit and the system memory, the hardware accelerator configured to perform at least a portion of an MBIR process on computer tomography data. The hardware accelerator unit may include one or more voxel evaluation modules which evaluate an updated value of a voxel given a voxel location in a reconstructed volume. By processing voxel data for voxels in a voxel neighborhood, processing time is reduces.

Abstract: A method includes obtaining spectral computed tomography (CT) information via an acquisition unit having an X-ray source and a CT detector. The method also includes, generating, with one or more processing units, using at least one image transform, a first basis image and a second basis image using the spectral CT information. Further, the method includes performing, with the one or more processing units, guided processing on the second basis image using the first basis image as a guide to provide a modified second basis image. Also, the method includes performing at least one inverse image transform using the first basis image and the modified second basis image to generate at least one modified image.

Abstract: A method of creating a hierarchical dictionary comprises, with a processor, extracting a number of symbols from a first image, constructing a number of refinement dictionary entries based on the symbols, the refinement dictionary entries forming a refinement dictionary, grouping a number of the refinement dictionary entries into clusters to form a number of refinement dictionary entry clusters, and constructing a number of direct dictionary entries for each of the refinement dictionary entry clusters, the direct dictionary entries forming a direct dictionary.