Abstract: The present disclosure provides a method for generating a backbone network, an apparatus for generating a backbone network, a device, and a storage medium. The method includes: acquiring a set of a training image, a set of an inference image, and a set of an initial backbone network; training and inferring, for each initial backbone network in the set of the initial backbone network, the initial backbone network by using the set of the training image and the set of the inference image, to obtain an inference time and an inference accuracy of a trained backbone network in an inference process; determining a basic backbone network based on the inference time and the inference accuracy of the trained backbone network in the inference process; and obtaining a target backbone network based on the basic backbone network and a preset target network.

Abstract: A video is segmented into a plurality of sequences corresponding to different facial states performed by a patient in the video. For each sequence, displacement of a plurality of groups of landmarks of a face of the patient is tracked, to obtain, for each group of the plurality of groups, one or more displacement measures characterizing positions of the landmarks of the group. The one or more displacement measures corresponding to each group are provided into a corresponding neural network, to obtain a landmark feature. The neural networks corresponding to each group are different from one another. A sequence score for the sequence is determined based on a plurality of landmark features corresponding to the groups. A plurality of sequence scores are provided into a machine learning component, to obtain a patient score. A disease state of the patient is determined based on the patient score.

Abstract: A computer-implemented method of performing deep learning based isocenter positioning includes acquiring a plurality of slabs covering an anatomical area of interest that comprises a patient's heart. For each slab, one or more deep learning models are used to determine a likelihood score for the slab indicating a probability that the slab includes at least a portion of the patient's heart. A center position of the patient's heart may then be determined based on the likelihood scores determined for the plurality of slabs.

Abstract: A method includes providing one or more instructions to perform a first action associated with administration of a medication; obtaining first data representing a capture of a patient performing the first action based on the one or more instructions as output by the output device; extracting information about movements of the patient from the first data; presenting an image configured to induce an emotional response in the patient; obtaining second data representing capture of the patient performing one or more microexpressions in response to presentation of the image; obtaining additional data including one or more of demographic information of the patient, information of a disease associated with the patient, or one or more characteristics of the medication; and based on the extracted information about the movements, the one or more microexpressions, and the additional data, determining a medical outcome including a progression of the disease in the patient.

Abstract: A method includes providing one or more instructions to perform a first action associated with administration of a medication; obtaining first data representing a capture of a patient performing the first action based on the one or more instructions; extracting information about movements of the patient from the first data; presenting an image configured to induce an emotional response in the patient; obtaining second data representing capture of the patient performing one or more microexpressions within 250 milliseconds of presentation of the image; obtaining additional data including one or more of demographic information of the patient, information of a disease associated with the patient, or one or more characteristics of the medication; and based on the extracted information about the movements, the one or more microexpressions, and the additional data, determining a medical outcome including a progression of the disease in the patient.

Abstract: Methods and systems for artificial intelligence based medical image segmentation are disclosed. In a method for autonomous artificial intelligence based medical image segmentation, a medical image of a patient is received. A current segmentation context is automatically determined based on the medical image and at least one segmentation algorithm is automatically selected from a plurality of segmentation algorithms based on the current segmentation context. A target anatomical structure is segmented in the medical image using the selected at least one segmentation algorithm.

Abstract: A method for real-time assessment of image information employing a computer that includes one or more processors includes obtaining, via a scanner, a magnetic resonance (MR) image of a region-of-interest, and providing data corresponding to at least a portion of the MR image to a deep learning model, where the deep learning model is previously trained based on one or more sets of training data. The method further includes assessing the data using the deep learning model and data obtained from an image quality database to obtain an assessed image quality value, and formulating an image quality classification in response to the assessed image quality value. The method additionally includes outputting the image quality classification within a predetermined time period from an initial scan of the region-of-interest.

Abstract: Methods and systems for artificial intelligence based medical image segmentation are disclosed. In a method for autonomous artificial intelligence based medical image segmentation, a medical image of a patient is received. A current segmentation context is automatically determined based on the medical image and at least one segmentation algorithm is automatically selected from a plurality of segmentation algorithms based on the current segmentation context. A target anatomical structure is segmented in the medical image using the selected at least one segmentation algorithm.

Abstract: Methods and systems for artificial intelligence based medical image segmentation are disclosed. In a method for autonomous artificial intelligence based medical image segmentation, a medical image of a patient is received. A current segmentation context is automatically determined based on the medical image and at least one segmentation algorithm is automatically selected from a plurality of segmentation algorithms based on the current segmentation context. A target anatomical structure is segmented in the medical image using the selected at least one segmentation algorithm.

Abstract: A system and method includes generation of one or more motion-corrupted images based on each of a plurality of reference images, and training of a regression network to determine a motion score, where training of the regression network includes input of a generated motion-corrupted image to the regression network, reception of a first motion score output by the regression network in response to the input image, and determination of a loss by comparison of the first motion score to a target motion score, the target motion score calculated based on the input motion-corrupted image and a reference image based on which the motion-corrupted image was generated.

Abstract: A learning-based magnetic resonance fingerprinting (MRF) reconstruction method for reconstructing an MR image of a tissue space in an MR scan subject for a particular MR sequence is disclosed. The method involves using a machine-learning algorithm that has been trained to generate a set of tissue parameters from acquired MR signal evolution without using a dictionary or dictionary matching.

Abstract: For heart segmentation in magnetic resonance or other medical imaging, deep learning trains a neural network. The neural network, such as U-net, includes at least one long-short-term memory (LSTM), such as a convolutional LSTM. The LSTM incorporates the temporal characteristics with the spatial to improve accuracy of the segmentation by the machine-learnt network.

Abstract: For heart segmentation in magnetic resonance or other medical imaging, deep learning trains a neural network. The neural network, such as U-net, includes at least one long-short-term memory (LSTM), such as a convolutional LSTM. The LSTM incorporates the temporal characteristics with the spatial to improve accuracy of the segmentation by the machine-learnt network.

Abstract: A method for real-time assessment of image information employing a computer that includes one or more processors includes obtaining, via a scanner, a magnetic resonance (MR) image of a region-of-interest, and providing data corresponding to at least a portion of the MR image to a deep learning model, where the deep learning model is previously trained based on one or more sets of training data. The method further includes assessing the data using the deep learning model and data obtained from an image quality database to obtain an assessed image quality value, and formulating an image quality classification in response to the assessed image quality value. The method additionally includes outputting the image quality classification within a predetermined time period from an initial scan of the region-of-interest.

Abstract: Methods and systems for artificial intelligence based medical image segmentation are disclosed. In a method for autonomous artificial intelligence based medical image segmentation, a medical image of a patient is received. A current segmentation context is automatically determined based on the medical image and at least one segmentation algorithm is automatically selected from a plurality of segmentation algorithms based on the current segmentation context. A target anatomical structure is segmented in the medical image using the selected at least one segmentation algorithm.

Abstract: A method for computing global longitudinal strain from cine magnetic resonance (MR) images includes automatically detecting landmark points in each MR long axis image frame included in a cine MR image sequence. A deformation field is determined between every pair of frames based on the landmark points. Myocardial pixels in the frames are labeled using a deep learning framework to yield myocardium masks. These myocardium masks are propagated to each frame using the deformation fields. A polar transformation is performed on each of the masked frames. The contours of the myocardium in each transformed frame are computed using a shortest path algorithm. Next, longitudinal strain is calculated at every pixel in the myocardium for the polar frames using the contours of the myocardium. Then, global longitudinal strain is computed by averaging the longitudinal strain at every pixel in the myocardium of the transformed frames.

Abstract: A method for computing global longitudinal strain from cine magnetic resonance (MR) images includes automatically detecting landmark points in each MR long axis image frame included in a cine MR image sequence. A deformation field is determined between every pair of frames based on the landmark points. Myocardial pixels in the frames are labeled using a deep learning framework to yield myocardium masks. These myocardium masks are propagated to each frame using the deformation fields. A polar transformation is performed on each of the masked frames. The contours of the myocardium in each transformed frame are computed using a shortest path algorithm. Next, longitudinal strain is calculated at every pixel in the myocardium for the polar frames using the contours of the myocardium. Then, global longitudinal strain is computed by averaging the longitudinal strain at every pixel in the myocardium of the transformed frames.

Abstract: Systems and methods for determining optimized imaging parameters for imaging a patient include learning a model of a relationship between known imaging parameters and a quality measure, the known imaging parameters and the quality measure being determined from training data. Optimized imaging parameters are determined by optimizing the quality measure using the learned model. Images of the patient are acquired using the optimized imaging parameters.

Abstract: A system and method includes generation of one or more motion-corrupted images based on each of a plurality of reference images, and training of a regression network to determine a motion score, where training of the regression network includes input of a generated motion-corrupted image to the regression network, reception of a first motion score output by the regression network in response to the input image, and determination of a loss by comparison of the first motion score to a target motion score, the target motion score calculated based on the input motion-corrupted image and a reference image based on which the motion-corrupted image was generated.

Abstract: A computer-implemented method of performing deep learning based isocenter positioning includes acquiring a plurality of slabs covering an anatomical area of interest that comprises a patient's heart. For each slab, one or more deep learning models are used to determine a likelihood score for the slab indicating a probability that the slab includes at least a portion of the patient's heart. A center position of the patient's heart may then be determined based on the likelihood scores determined for the plurality of slabs.