Canon Medical Systems Corporation

Toshiba Medical Systems Corporation

University Health Network

Ting Xia

Zhou Yu

Jian Zhou

Liang Cai

Wenli Wang

Evren Asma

Wenyuan Qi

Bernice Hoppel

Ilmar A Hein

Yujie Lu

Masakazu Matsuura

Patrik Rogalla

An information processing method of an embodiment is a processing method of information acquired by imaging performed by a medical image diagnostic apparatus, the information processing method includes the steps of: acquiring noise data by imaging a phantom using a medical imaging apparatus; based on first subject projection data acquired by the imaging performed by a medical image diagnostic modality of a same kind as the medical image diagnostic apparatus and the noise data, acquiring synthesized subject data in which noise based on the noise data is added to the first subject projection data; and acquiring a noise reduction processing model by machine learning using the synthesized subject data and second subject projection data acquired by the imaging performed by the medical image diagnostic modality.

Information processing method, medical image diagnostic apparatus, and information processing system

A method, apparatus, and computer-readable storage medium for controlling exposure/irradiation during a main three-dimensional X-ray imaging scan using at least one spatially-distributed characteristic of a pre-scan/scout scan preceding the main scan. The at least one spatially-distributed characteristic includes (1) a spatially-distributed noise characteristic of the pre-scan and/or (2) a spatially-distributed identification of exposure-sensitive tissue types. The at least one spatially-distributed characteristic can be calculated from images reconstructed from sinogram/projection data and/or from sinogram/projection directly using a neural network.

Apparatus and methods for image quality improvement based on noise and dose optimization

The present disclosure relates to a method for patient-specific optimization of imaging protocols. According to an embodiment, the present disclosure relates to a method for generating a patient-specific imaging protocol, comprising acquiring scout scan data, the scout scan data including scout scan information and scout scan parameters, generating a simulated image based on the acquired scout scan data, deriving a simulated dose map from the generated simulated image, determining image quality of the generated simulated image by applying machine learning to the generated simulated image, the neural network being trained to generate at least one probabilistic quality representation corresponding to at least one region of the generated simulated image, evaluating the determined image quality relative to a image quality threshold and the derived simulated dose map relative to a dosage threshold, optimizing. based on the evaluating, scan acquisition parameters and image reconstruction parameters, and generating, optimal imaging protocol parameters, wherein the optimal imaging protocol parameters maximize image quality while minimizing radiation exposure.

System and methods for radiographic image quality assessment and protocol optimization

A method and apparatus are provided that use deep learning (DL) networks to reduce noise and artifacts in reconstructed computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI) images. DL networks are used in both the sinogram and image domains. In each domain, a detection network is used to (i) determine if particular types of artifacts are exhibited (e.g., beam-hardening artifact, ring, motion, metal, photon-starvation, windmill, zebra, partial-volume, cupping, truncation, streak artifact, and/or shadowing artifacts), (ii) determine whether the detected artifact can be corrected through a changed scan protocol or image-processing techniques, and (iii) determine whether the detected artifacts are fatal, in which case the scan is stopped short of completion. When the artifacts can be corrected, corrective measures are taken through a changed scan protocol or through image processing to reduce the artifacts (e.g., convolutional neural network can be trained to perform the image processing).

Apparatus and method for artifact detection and correction using deep learning

A method and apparatus is provided to obtain projection data representing an intensity of X-ray radiation detected at a plurality of detector elements after traversing an object, the projection data being corrected for a baseline offset, correct the projection data by performing a positivity mapping to generate corrected projection data, perform a logarithm operation on the corrected projection data to generate post-log projection data, correct for a bias of the post-log projection data, using the projection data, to generate bias-corrected projection data, and reconstruct an image of the object from the bias-corrected projection data.

Apparatus and method for correcting bias in low-count computed tomography projection data

An apparatus for performing a non-local means (NLM) filter is described. The pixel of the NLM-filtered image are weighted averages of pixels from a noisy image, where the weights are a measure of the similarity between patches of the noisy image. The similarity weights can be calculated using a Kullback-Leibler or a Euclidean distance measure. The similarity weights can be based on filtered patches of the noisy image. The similarity weights can be based on a similarity measure between patches of an anatomical image corresponding to the noisy image. The similarity weights can be calculated using a time series of noisy images to increase the statistical sample size of the patches. The similarity weights can be calculated using a weighted sum of channel similarity weights calculated between patches of noisy image that have been band-pass filtered. The NLM-filtered image can also be blended with a non-NLM-filtered image.

Growing community of inventors

Mundelein, IL, United States of America

Ting Xia