Abstract: Deep learning based systems, methods, and instrumentalities are described herein for registering images from a same imaging modality and different imaging modalities. Transformation parameters associated with the image registration task are determined using a neural ordinary differential equation (ODE) network that comprises multiple layers, each configured to determine a respective gradient update for the transformation parameters based on a current state of the transformation parameters received by the layer. The gradient updates determined by the multiple ODE layers are then integrated and applied to initial values of the transformation parameters to obtain final parameters for completing the image registration task. The operations of the ODE network may be facilitated by a feature extraction network pre-trained to determine content features shared by the input images.

Abstract: A system comprising a first computing apparatus in communication with multiple second computing apparatuses. The first computing apparatus may obtain a plurality of first trained machine learning models for a task from the multiple second computing apparatuses. At least a portion of parameter values of the plurality of first trained machine learning models may be different from each other. The first computing apparatus may also obtain a plurality of training samples. The first computing apparatus may further determine, based on the plurality of training samples, a second trained machine learning model by learning from the plurality of first trained machine learning models.

Abstract: A method for lung nodule evaluation is provided. The method may include obtaining a target image including at least a portion of a lung of a subject. The method may also include segmenting, from the target image, at least one target region each of which corresponds to a lung nodule of the subject. The method may further include generating an evaluation result with respect to the at least one lung nodule based on the at least one target region.

Abstract: System for image correction in PET is provided. The system may acquire a PET image and a CT image of a subject. The system may generate, based on the PET image and the CT image, an attenuation-corrected PET image of the subject by application of an attenuation correction model. The attenuation correction model may be a trained cascaded neural network including a trained first model and at least one trained second model downstream to the trained first model. During the application of the attenuation correction model, an input of each of the at least one trained second model may include the PET image, the CT image, and an output image of a previous trained model that is upstream and connected to the trained second model.

Abstract: A system for PET image reconstruction is provided. The system may obtain PET data of a subject. The PET data may be associated with a plurality of coincidence events, which includes scattering events. The system may also generate a preliminary scatter sinogram relating to the scattering events based on the PET data. The system may also generate a target scatter sinogram relating to the scattering events by applying a scatter sinogram generator based on the preliminary scatter sinogram. The target scatter sinogram may have a higher image quality than the preliminary scatter sinogram. The system may further reconstruct a target PET image of the subject based on the PET data and the target scatter sinogram.

Abstract: Healthcare services can be automated utilizing a system that recognizes at least one characteristic of a patient based on images of the patient acquired by an image capturing device. Relying on information extracted from these images, the system may automate multiple aspects of a medical procedure such as patient identification and verification, positioning, diagnosis and/or treatment planning using artificial intelligence or machine learning techniques. By automating these operations, healthcare services can be provided remotely and/or with minimum physical contact between the patient and a medical professional.

Abstract: A system for Magnetic Resonance Imaging (MRI) is provided. The system may obtain at least one training sample each of which includes full MRI data. The system may also obtain a preliminary subsampling model and a preliminary MRI reconstruction model. The system may further generate a subsampling model corresponding to an MRI reconstruction model by jointly training the preliminary subsampling model and the preliminary MRI reconstruction model using the at least one training sample. The subsampling model may be the trained preliminary subsampling model, and the MRI reconstruction model may be at least a portion of the trained preliminary MRI reconstruction model.

Abstract: Image-based key points detection using a convolutional neural network (CNN) may be impacted if the key points are occluded in the image. Images obtained from additional imaging modalities such as depth and/or thermal images may be used in conjunction with RGB images to reduce or minimize the impact of the occlusion. The additional images may be used to determine adjustment values that are then applied to the weights of the CNN so that the convolution operations may be performed in a modality aware manner to increase the robustness, accuracy, and efficiency of key point detection.

Abstract: Disclosed herein are systems, methods, and instrumentalities associated with reconstructing magnetic resonance (MR) images based on under-sampled MR data. The MR data include 2D or 3D information, and may encompass multiple contrasts and multiple coils. The MR images are reconstructed using deep learning (DL) methods, which may accelerate the scan and/or image generation process. Challenges imposed by the large quantity of the MR data and hardware limitations are overcome by separately reconstructing MR images based on respective subsets of contrasts, coils, and/or readout segments, and then combining the reconstructed MR images to obtain desired multi-contrast results.

Abstract: Neural network based systems, methods, and instrumentalities may be used to remove motion artifacts from magnetic resonance (MR) images. Such a neural network based system may be trained to perform the motion artifact removal tasks without reference (e.g., without using paired motion-contaminated and motion-free MR images). Various training techniques are described herein including one that feeds the neural network with pairs of MR images with different levels of motion contamination and forces the neural network learn to correct the motion contamination by transforming a first image of a contaminated pair into a second image of the contaminated pair. Other neural network training techniques are also described with an aim to reduce the reliance on training data that is difficult to obtain.

Abstract: A method for automated calibration is provided. The method may include obtaining a plurality of interest points based on prior information regarding a device and image data of the device captured by a visual sensor. The method may include identifying at least a portion of the plurality of interest points from the image data of the device. The method may also include determining a transformation relationship between a first coordinate system and a second coordinate system based on information of at least a portion of the identified interest points in the first coordinate system and in the second coordinate system that is applied to the visual sensor or the image data of the device.

Abstract: The pose and shape of a human body may be recovered based on joint location information associated with the human body. The joint location information may be derived based on an image of the human body or from an output of a human motion capture system. The recovery of the pose and shape of the human body may be performed by a computer-implemented artificial neural network (ANN) trained to perform the recovery task using training datasets that include paired joint location information and human model parameters. The training of the ANN may be conducted in accordance with multiple constraints designed to improve the accuracy of the recovery and by artificially manipulating the training data so that the ANN can learn to recover the pose and shape of the human body even with partially observed joint locations.

Abstract: Methods and systems for classifying an image.

Abstract: Systems, methods and instrumentalities are described herein for automating a medical environment. The automation may be realized using one or more sensing devices and at least one processing device. The sensing devices may be configured to capture images of the medical environment and provide the images to the processing device. The processing device may determine characteristics of the medical environment based on the images and automate one or more aspects of the operations in the medical environment. These characteristics may include, e.g., people and/or objects present in the images and respective locations of the people and/or objects in the medical environment. The operations that may be automated may include, e.g., maneuvering and/or positioning a medical device based on the location of a patient, determining and/or adjusting the parameters of a medical device, managing a workflow, providing instructions and/or alerts to a patient or a physician, etc.

Abstract: Motion contaminated magnetic resonance (MR) images for training an artificial neural network to remove motion artifacts from the MR images are difficult to obtain. Described herein are systems, methods, and instrumentalities for injecting motion artifacts into clean MR images and using the artificially contaminated images for machine learning and neural network training. The motion contaminated MR images may be created based on clean source MR images that are associated with multiple physiological cycles of a scanned object, and by deriving MR data segments for the multiple physiological cycles based on the source MR images. The MR data segments thus derived may be combined to obtain a simulated MR data set, from which one or more target MR images may be generated to exhibit a motion artifact. The motion artifact may be created by manipulating the source MR images and/or controlling the manner in which the MR data set or the target MR images are generated.

Abstract: A bullseyes plot may be generated based on cardiac magnetic resonance imaging (CMRI) to facilitate the diagnosis and treatment of heart diseases. Described herein are systems, methods, and instrumentalities associated with bullseyes plot generation. A plurality of myocardial segments may be obtained for constructing the bullseye plot based on landmark points detected in short-axis and long-axis magnetic resonance (MR) slices of the heart and by arranging the short-axis MR slices sequentially in accordance with the order in which the slices are generated during the CMRI. The sequential order of the short-axis MR slices may be determined utilizing projected locations of the short-axis MR slices on a long-axis MR slice and respective distances of the projected locations to a landmark point of the long-axis MR slice. The myocardium and/or landmark points may be identified in the short-axis and/or long-axis MR slices using artificial neural networks.

Abstract: Described herein are neural network-based systems, methods and instrumentalities associated with estimating a thickness of an anatomical structure based on a visual representation of the anatomical structure and a machine-learned thickness prediction model. The visual representation may include an image or a segmentation mask of the anatomical structure. The thickness prediction model may be learned based on ground truth information derived by applying a partial differential equation such as Laplace's equation to the visual representation and solving the partial differential equation. When the visual representation includes an image of the anatomical structure, the systems, methods and instrumentalities described herein may also be capable of generating a segmentation mask of the anatomical structure based on the image.

Abstract: An unsupervised machine learning method with self-supervision losses improves a slice-wise spatial resolution of 3D medical images with thick slices, and does not require high resolution images as the ground truth for training. The method utilizes information from high-resolution dimensions to increase a resolution of another desired dimension.

Abstract: System for image correction in PET is provided. The system may acquire a PET image and a CT image of a subject. The system may generate, based on the PET image and the CT image, an attenuation-corrected PET image of the subject by application of an attenuation correction model. The attenuation correction model may be a trained cascaded neural network including a trained first model and at least one trained second model downstream to the trained first model. During the application of the attenuation correction model, an input of each of the at least one trained second model may include the PET image, the CT image, and an output image of a previous trained model that is upstream and connected to the trained second model.

Abstract: The present disclosure provides a system for image reconstruction. The system may obtain an initial image of a subject. The initial image may be generated based on scan data of the subject that is collected by an imaging device. The system may also generate a gradient image associated with the initial image. The system may further generate a target image of the subject by applying an image reconstruction model based on the initial image and the gradient image. The target image may have a higher image quality than the initial image.