Abstract: Embodiments disclosed herein generally relate to representative datasets for biomedical machine learning models. Particularly, aspects of the present disclosure are directed to identifying a representative distribution of characteristics for a disease, generating a dataset comprising a set of biomedical images, wherein the dataset has a distribution of the characteristics that corresponds to the representative distribution of the characteristics for the disease, processing the dataset using a trained machine learning model, and outputting a result of the processing, wherein the result corresponds to a prediction that a biomedical image of the dataset includes a depiction of a set of tumor cells or other structural and/or functional biological entities associated with the disease, the biomedical image is associated with a diagnosis of the disease, the biomedical image is associated with a classification of the disease, and/or the biomedical image is associated with a prognosis for the disease.

Abstract: The present disclosure relates to techniques for efficient development of initial models and efficient model update and/or adaptation to a different image domain using an adaptive learning framework. For efficient development of initial models, a two-step development strategy may be performed as follows: Phase 1: Model preconditioning, where an artificial intelligence system leverages existing annotated datasets and improves learning skills through training of these datasets; and Phase 2: Target-model training, where an artificial intelligence system utilizes the learning skills learned from Phase 1 to extend itself to a different image domain (target domain) with less number of annotations required in the target domain than conventional learning methods.

Abstract: A computer implemented method for identifying at least one peak in a mass spectrometry response curve is provided comprising: a) providing at least one mass spectrometry response curve by using at least one mass spectrometry device; b) evaluating the mass spectrometry response curve by using at least one trained model thereby identifying a start point and an end point of at least one peak of the mass spectrometry response curve, wherein the model was trained using a deep learning regression architecture.

Abstract: An image analysis system and method classify cells in a tissue image. The system and method may extract at least one image feature characterizing an object in the tissue image. Based on the extracted image feature, cells may be classified according to at least one predefined membrane pattern. For each classified cell, a class label that identifies a class to which the classified cell belongs may be outputted.

Abstract: An image analysis system and method to generate selective stain segmentation images for at least one cell type of interest within a stained tissue image are presented. Cell membranes in the tissue image are detected to generate a corresponding membrane mask image. Cell nuclei are classified to generate a classified cells image. Selective stain segmentation images are generated for the at least one cell type of interest based on the membrane mask image and the classified cells image.

Abstract: A system and a method for unveiling poorly visible or lightly colored nuclei in an input image are disclosed. An input image is fed to a color deconvolution module for deconvolution into two color channels that are processed separately before being combined. The input image is deconvolved into two separate images: a stain image and a counter stain image. A complement of the stain image is generated in order to clearly reflect the locations of the poorly visible or light-colored nuclei. The complement image and the counter stain image are optionally normalized and then combined and segmented, to generate an output image with clearly defined nuclei. Alternatively, the complement of the stain image and the counter stain image can optionally be normalized, and then segmented prior to being combined to generate the output image.

Abstract: A digital pathology system and associated method and computer program product provide a quantitative analysis of entire tissue slides as well as intuitive, effective, fast, and precise quantification of biomarker expressions across relevant areas of the entire tissue slides. The digital pathology system enables a novel workflow that allows the user to efficiently outline clinically relevant morphology in its entirety, including solid tumor areas. Quantitative analysis results are then efficiently and intuitively provided to the user for all tissue content (i.e., millions of cells) within seconds. This efficiency is made possible by a pre-computation step that computes and stores image analysis results for later retrieval. Visualizing vast amount of data effectively is another component of the system that provides information and confidence to the user about the biomarker expression levels.

Abstract: An image analysis system and method to generate selective stain segmentation images for at least one cell type of interest within a stained tissue image are presented. Cell membranes in the tissue image are detected to generate a corresponding membrane mask image. Cell nuclei are classified to generate a classified cells image. Selective stain segmentation images are generated for the at least one cell type of interest based on the membrane mask image and the classified cells image.

Abstract: An image analysis system and method classify cells in a tissue image. The system and method may extract at least one image feature characterizing an object in the tissue image. Based on the extracted image feature, cells may be classified according to at least one predefined membrane pattern. For each classified cell, a class label that identifies a class to which the classified cell belongs may be outputted.

Abstract: A system and a method for unveiling poorly visible or lightly colored nuclei in an input image are disclosed. An input image is fed to a color deconvolution module for deconvolution into two color channels that are processed separately before being combined. The input image is deconvolved into two separate images: a stain image and a counter stain image. A complement of the stain image is generated in order to clearly reflect the locations of the poorly visible or light-colored nuclei. The complement image and the counter stain image are optionally normalized and then combined and segmented, to generate an output image with clearly defined nuclei. Alternatively, the complement of the stain image and the counter stain image can optionally be normalized, and then segmented prior to being combined to generate the output image.

Abstract: A digital pathology system and associated method and computer program product provide a quantitative analysis of entire tissue slides as well as intuitive, effective, fast, and precise quantification of biomarker expressions across relevant areas of the entire tissue slides. The digital pathology system enables a novel workflow that allows the user to efficiently outline clinically relevant morphology in its entirety, including solid tumor areas. Quantitative analysis results are then efficiently and intuitively provided to the user for all tissue content (i.e., millions of cells) within seconds. This efficiency is made possible by a pre-computation step that computes and stores image analysis results for later retrieval. Visualizing vast amount of data effectively is another component of the system that provides information and confidence to the user about the biomarker expression levels.

Abstract: Phase unwrapping is provided for phase contrast magnetic resonance (MR) imaging. The velocity values are unaliased. For a given location over time, a path over time through a directed graph of possible velocities at each time is determined by minimization of derivatives over time. The possible velocities are based on the input velocity, the input velocity wrapped in a positive direction, and the input velocity wrapped in a negative direction, so the selection to create the minimum cost path represents unaliasing of any aliased velocities.

Abstract: A method of magnetic resonance (MR) imaging of a volume undergoing repetitive motion includes obtaining source slice data indicative of a plurality of source slices during the repetitive motion, and obtaining anchor slice data indicative of an anchor slice during the repetitive motion. The anchor slice intersects the plurality of source slices. The source slice data and the anchor slice data are reconstructed. A three-dimensional image assembly procedure is implemented to generate, for each phase of the repetitive motion, volume data based on a respective subset of the reconstructed source slice data. For each phase of the repetitive motion, the respective subset of slices is selected based on a correlation of the source slice data and the anchor slice data along an intersection between each source slice and the anchor slice. The source slice data of the selected subset is corrected for misalignment with the anchor slice data.

Abstract: Magnetic resonance reconstruction includes motion compensation. Inverse-consistent non-rigid registration is used to determine motion between shots. The motion is incorporated into reconstruction. The incorporation compensates for the motion resulting from the period over which the MR data is acquired.

Abstract: A method of magnetic resonance (MR) imaging of a volume undergoing repetitive motion includes obtaining source slice data indicative of a plurality of source slices during the repetitive motion, and obtaining anchor slice data indicative of an anchor slice during the repetitive motion. The anchor slice intersects the plurality of source slices. The source slice data and the anchor slice data are reconstructed. A three-dimensional image assembly procedure is implemented to generate, for each phase of the repetitive motion, volume data based on a respective subset of the reconstructed source slice data. For each phase of the repetitive motion, the respective subset of slices is selected based on a correlation of the source slice data and the anchor slice data along an intersection between each source slice and the anchor slice. The source slice data of the selected subset is corrected for misalignment with the anchor slice data.

Abstract: A method for performing motion compensation in a series of magnetic resonance (MR) images includes acquiring a set of MR image frames spanning different points along an MR recovery curve. A motion-free synthetic image is generated for each of the acquired MR image frames using prior knowledge pertaining to an MR recovery curve. Each of the acquired MR images is registered to its corresponding generated synthetic images. Motion within each of the acquired MR image is corrected based on its corresponding generated synthetic image that has been registered thereto.

Abstract: A method including receiving an image sequence, wherein the image sequence includes a plurality of two-dimensional (2D) image frames of an organ arranged in a time sequence; constructing a three-dimensional (3D) volume by stacking a plurality of the 2D image frames in time order; detecting a best bounding box for a target of interest in the 3D volume, wherein the best bounding box is specified by a plurality of parameters including spatial and temporal information contained in the 3D volume; and determining the target of interest from the best bounding box.

Abstract: A method for identifying a region of interest within a time sequence of images includes acquiring a time sequence of images comprising a plurality of image frames. Image segmentation is performed to segment a region of interest (ROI) from within each of the plurality of image frames of the time sequence of images. Manual edits are received for the ROI within one or more of the plurality of image frames. The manual edits are propagated to other image frames of the plurality of images. An extent to which each of the manual edits are propagated to other image frames is dependent upon a transformation function or deformation field used to propagate the manual edits and a weighing factor that is influenced by a distance in time between the other image frames and the frames that have been manually edited.

Abstract: A method for clinical parameter derivation and adaptive flow acquisition within a sequence of magnetic resonance images includes commencing an acquisition of a sequence of images. One or more landmarks are automatically detected from within one or more images of the sequence of images. The detected one or more landmarks are propagated across subsequent images of the sequence of images. A plane is fitted to the propagation of landmarks. The positions of landmarks or alternatively the position of the fitted plane within the sequence of images is used for derivation of clinical parameters such as tissue velocities and/or performing adaptive flow acquisitions to measure blood flow properties.

Abstract: A system receives cardiac cine MR images consists of multiple slices of the heart over time. A series of short axis images slices are received. Long axis images are also received by the system, wherein a base plane defined by landmark points is detected. An intersection of the base plane with a contour of a heart chamber is determined for a plurality of slices in the short axis image. A volume for each of the contour slices covering the heart chamber, including for contours that are limited by base plane intersections, is evaluated. All slice volumes are summed to determine a total volume of the chamber. In one embodiment the chamber is a left ventricle and the landmark is a mitral valve. An ejection factor is determined.