Abstract: The disclosure relates to systems and methods for evaluating a blood vessel. The method includes receiving image data of the blood vessel acquired by an image acquisition device, and predicting, by a processor, blood vessel condition parameters of the blood vessel by applying a deep learning model to the acquired image data of the blood vessel. The deep learning model maps a sequence of image patches on the blood vessel to blood vessel condition parameters on the blood vessel, where in the mapping the entire sequence of image patches contribute to the blood vessel condition parameters. The method further includes providing the blood vessel condition parameters of the blood vessel for evaluating the blood vessel.

Abstract: A computer-implemented method for automatically detecting a target object from a 3D image is disclosed. The method may include receiving the 3D image acquired by an imaging device. The method may further include detecting, by a processor, a plurality of bounding boxes as containing the target object using a 3D learning network. The learning network may be trained to generate a plurality of feature maps of varying scales based on the 3D image. The method may also include determining, by the processor, a set of parameters identifying each detected bounding box using the 3D learning network, and locating, by the processor, the target object based on the set of parameters.

Abstract: Embodiments of the disclosure provide systems and methods for detecting cancer metastasis in a whole-slide image. The system may include a communication interface configured to receive the whole-slide image and a learning model. The whole-slide image is acquired by an image acquisition device. The system may also include a memory configured to store a plurality of tiles derived from the whole-slide image in a queue. The system may further include at least one processor, configured to apply the learning model to at least two tiles stored in the queue in parallel to obtain detection maps each corresponding to a tile, and detect the cancer metastasis based on the detection maps.

Abstract: The disclosure relates to systems and methods for determining blood vessel conditions. The method includes receiving a sequence of image patches along a blood vessel path acquired by an image acquisition device. The method also includes predicting a sequence of blood vessel condition parameters on the blood vessel path by applying a trained deep learning model to the acquired sequence of image patches on the blood vessel path. The deep learning model includes a data flow neural network, a recursive neural network and a conditional random field model connected in series. The method further includes determining the blood vessel condition based on the sequence of blood vessel condition parameters. The disclosed systems and methods improve the calculation of the sequence of blood vessel condition parameters through an end-to-end training model, including improving the calculation speed, reducing manual intervention for feature extraction, increasing accuracy, and the like.

Abstract: Methods and Systems for generating a centerline for an object in an image and computer readable medium are provided. The method includes receiving an image containing the object. The method also includes generating the centerline of the object by tracing a sequence of patches with a virtual agent. For each patch other than the initial patch, the method determines a current patch based on the position and action of the virtual agent at a previous patch. The method further determines a policy function and a value function based on the current patch using a trained learning network, which includes an encoder followed by a first learning network and a second learning network. The learning network is trained by maximizing a cumulative reward. The method also determines the action of the virtual agent at the current patch. Additionally, the method displays the centerline of the object.

Abstract: The disclosure provides a method and device for performing three-dimensional blood vessel reconstruction using projection images of a patient. The computer-implemented method includes receiving a first two-dimensional image of a blood vessel in a first projection direction and a three-dimensional model of the blood vessel. The method further includes determining, by a processor, a first optical path length at a selected position of the blood vessel based on the first two-dimensional image. The method also includes determining, by the processor, a second optical path length at the selected position of the blood vessel in the three-dimensional model. The method additional includes adjusting the three-dimensional model of the blood vessel, based on a comparison of the first optical path length and the second optical path length.

Abstract: The disclosure provides a method and device for performing three-dimensional blood vessel reconstruction using angiographic images. The method includes an acquisition step that acquires a first two-dimensional image of the blood vessel in the first projection direction and a corresponding reconstructed three-dimensional model of the blood vessel.

Abstract: The present disclosure relates to a device, a system, and a computer-readable medium for calculating vessel flow parameters based on angiography. In one implementation, the device includes a processor and a memory storing computer-executable instructions that, when executed by the processor, cause the processor to perform the following operations: selecting a plurality of template frames from the angiographic images to generate a 3D model for a vessel; determining a start frame and an end frame in the plurality of angiographic images showing a contrast filling process; determining corresponding locations of front ends of the contrast in the start frame and the end frame in the 3D model of the vessel; calculating a vessel volume between the determined locations of the front ends in the 3D model; and determining an average blood flow rate based on the calculated volume, and a time interval between the start frame and the end frame.

Abstract: Embodiments of the disclosure provide systems and methods for processing unstructured texts in a medical record. A disclosed system includes at least one processor configured to determine a plurality of word representations of an unstructured text and tag entities in the unstructured text by performing a named entity recognition task on the plurality of word representations. The at least one processor is further configured to determine position embeddings based on positions of words in the unstructured text relative to positions of the tagged entities and concatenate the plurality of word representations with the position embeddings. The at least one processor is also configured to determine relation labels between pairs of tagged entities by performing a relationship extraction task on the concatenated word representations and position embeddings.

Abstract: Systems and methods for generating a centerline for an object in an image are provided. An exemplary method includes receiving an image containing the object. The method also includes generating a distance cost image using a trained first learning network based on the image. The method further includes detecting end points of the object using a trained second learning network based on the image. Moreover, the method includes extracting the centerline of the object based on the distance cost image and the end points of the object.

Abstract: Embodiments of the disclosure provide systems and methods for analyzing a biomedical image including at least one tree structure object. The system includes a communication interface configured to receive a learning model and a plurality of model inputs derived from the biomedical image. The biomedical image is acquired by an image acquisition device. The system further includes at least one processor configured to apply the learning model to the plurality of model inputs to analyze the biomedical image. The learning model includes a first network configured to process the plurality of model inputs to construct respective feature maps and a second network configured to process the feature maps collectively. The second network is a tree structure network that models a spatial constraint of the tree structure object.

Abstract: The present disclosure is directed to a method and device for generating anatomical labels for a physiological tree structure. The method may include receiving a 3D model and a 3D skeleton line of the physiological tree structure. The 3D model is restructured based on medical image data of the physiological tree structure acquired by an imaging device. The method further includes selecting at least one level from extracting geometrical features from a pool of selectable levels. The method also includes extracting, by a processor, geometrical features from the 3D model of the physiological tree structure along the 3D skeleton line at the selected at least one level. The method also includes generating, by the processor, anatomical labels for the physiological tree structure using a trained learning network based on the extracted geometrical features.

Abstract: The present disclosure relates to a method, storage medium, and system for analyzing an image sequence of a periodic physiological activity. In one implementation, the method includes receiving the image sequence acquired by an imaging device, the image sequence having a plurality of frames and determining local motions for pixels in each frame of the image sequence. The local motion for a pixel may be determined using corresponding pixels in frames adjacent to the frame to which the pixel belongs. The method further includes determining principal motions for the plurality of frames based on the local motions; determining a motion magnitude profile based on the principal motions; and determining the phase of each frame in the image sequence based on the motion magnitude profile.

Abstract: The present disclosure is directed to a method and device for automatically predicting FFR based on images of vessel. The method for automatically predicting FFR based on images of a vessel. The method comprises a step of receiving the images of a vessel acquired by an imaging device. Then, a sequence of flow speeds at a sequence of positions on a centerline of the vessel is acquired by a processor. A sequence of first features at the sequence of positions on a centerline of the vessel are acquired by the processor, by fusing structure-related features and flow speeds and using a convolutional neural network. Then, a sequence of FFR at the sequence of positions is determined by the processor through using a sequence-to-sequence neural network on the basis of the sequence of first features.

Abstract: Embodiments of the disclosure provide systems and methods for analyzing a biomedical image including at least one tree structure object. The system includes a communication interface configured to receive a learning model and a plurality of model inputs derived from the biomedical image. The biomedical image is acquired by an image acquisition device. The system further includes at least one processor configured to apply the learning model to the plurality of model inputs to analyze the biomedical image. The learning model includes a first network configured to process the plurality of model inputs to construct respective feature maps and a second network configured to process the feature maps collectively. The second network is a tree structure network that models a spatial constraint of the tree structure object.

Abstract: Methods and systems for segmenting images having sparsely distributed objects are disclosed. A method may include: predicting object potential areas in the image using a preliminary fully convolutional neural network; segmenting a plurality of sub-images corresponding to the object potential areas in the image using a refinement fully convolutional neural network, wherein the refinement fully convolutional neural network is trained to segment images on a higher resolution compared to a lower resolution utilized by the preliminary fully convolutional neural network; and combining the segmented sub-images to generate a final segmented image.

Abstract: Embodiments of the disclosure provide systems and methods for segmenting a biomedical image including at least one tree structure object. The system includes a communication interface configured to receive the biomedical image and a learning model. The biomedical image is acquired by an image acquisition device. The system further includes at least one processor configured to extract a plurality of image patches from the biomedical image and apply the learning model to the plurality of image patches to segment the biomedical image. The learning model includes a convolutional network configured to process the plurality of image patches to construct respective feature maps and a tree structure network configured to process the feature maps collectively to obtain a segmentation mask for the tree structure object. The tree structure network models a spatial constraint of the plurality of image patches.

Abstract: Embodiments of the disclosure provide systems and methods for segmenting a biomedical image including at least one tree structure object. The system includes a communication interface configured to receive the biomedical image and a learning model. The biomedical image is acquired by an image acquisition device. The system further includes at least one processor configured to extract a plurality of image patches from the biomedical image and apply the learning model to the plurality of image patches to segment the biomedical image. The learning model includes a convolutional network configured to process the plurality of image patches to construct respective feature maps and a tree structure network configured to process the feature maps collectively to obtain a segmentation mask for the tree structure object. The tree structure network models a spatial constraint of the plurality of image patches.

Abstract: The disclosure provides a method and device for performing three-dimensional blood vessel reconstruction using angiographic images. The method includes an acquisition step that acquires a first two-dimensional image of the blood vessel in the first projection direction and a corresponding reconstructed three-dimensional model of the blood vessel.

Abstract: Embodiments of the disclosure provide systems and methods for detecting cancer metastasis in a whole-slide image. The system may include a communication interface configured to receive the whole-slide image and a learning model. The whole-slide image is acquired by an image acquisition device. The system may also include a memory configured to store a plurality of tiles derived from the whole-slide image in a queue. The system may further include at least one processor, configured to apply the learning model to at least two tiles stored in the queue in parallel to obtain detection maps each corresponding to a tile, and detect the cancer metastasis based on the detection maps.