Abstract: Systems and methods for generating a response summarizing patient data are provided. One or more prompts, comprising 1) patient data retrieved from one or more patient databases and 2) instructions, are received. A response summarizing the patient data is generated based on the instruction using a large language model. The response is output.

Abstract: Systems and methods for automatically determining an image quality assessment of a rendered medical image are provided. A rendered medical image is received. One or more measures of interest are extracted from the rendered medical image. An image quality assessment of the rendered medical image is determined using a machine learning based image quality assessment network based on the one or more measures of interest. The image quality assessment of the rendered medical image is output.

Abstract: For generating and/or machine training to generate a whole-body representation of a patient, one or more partial-body scans or images of the patient are extrapolated to the whole-body representation of both interior and exterior anatomy. One or more machine-learned models (e.g., per-organ-group implicit generative shape models) fill the whole-body representation based on the partial information from imaging.

Abstract: The present embodiment relates to a renderer and an interactive method for image editing of medical 3D anatomical data. The method includes receiving a dataset with volumetric image data, which have been acquired from an image acquisition modality, and providing a signed distance field data structure of the received dataset. Further, editing operations are received from a user interface for editing at least a part of the provided signed distance field data structure. A visualization of the editing operations is calculated and displayed on a display.

Abstract: A computer-implemented method for providing information from an electronic medical record database is provided. The method comprising the following steps of receiving a query for the information, identifying at least one code of a medical ontology based on the query, determining one or more related codes in the medical ontology based on the medical ontology and the identified at least one code, the one or more related codes having a relation to the at least one identified code in the medical ontology, parsing the electronic medical record database to identify entries in the electronic medical record database comprising one or more of the identified at least one code and the one or more related codes, and providing the information based on the identified entries.

Abstract: Systems and methods for automatically determining an image quality assessment of a rendered medical image are provided. A rendered medical image is received. One or more measures of interest are extracted from the rendered medical image. An image quality assessment of the rendered medical image is determined using a machine learning based image quality assessment network based on the one or more measures of interest. The image quality assessment of the rendered medical image is output.

Abstract: The present embodiment relates to a renderer and an interactive method for image editing of medical 3D anatomical data. The method includes receiving a dataset with volumetric image data, which have been acquired from an image acquisition modality, and providing a signed distance field data structure of the received dataset. Further, editing operations are received from a user interface for editing at least a part of the provided signed distance field data structure. A visualization of the editing operations is calculated and displayed on a display.

Abstract: A medical data processing system includes: a medical data storage configured to collect one or more pieces of medical data from one or more input devices, a data analyzer configured to execute one or more analyzing services on the one or more pieces of medical data so as to output one or more corresponding pieces of analyzed medical data, a results manager configured to collect usage data, related to how the one or more pieces of analyzed medical data are analyzed by the data analyzer and/or used by one or more users, and a processing optimizer configured to control operation of the data analyzer based on the usage data.

Abstract: Systems and methods are provided rendering a three-dimensional volume. Scan data representing an anatomical object of a patient is acquired. A boundary of the object is identified in the scan data. A first lightmap is positioned inside the boundary of the object. A second lightmap is positioned outside the boundary of the object. The three-dimensional volume of the object is rendered from the scan data with lighting based on the first lightmap and second lightmap.

Abstract: For three-dimensional rendering, a machine-learnt model is trained to generate representation vectors for rendered images formed with different rendering parameter settings. The distances between representation vectors of the images to a reference are used to select the rendered image and corresponding rendering parameters that provides a consistency with the reference. In an additional or different embodiment, optimized pseudo-random sequences are used for physically-based rendering. The random number generator seed is selected to improve the convergence speed of the renderer and to provide higher quality images, such as providing images more rapidly for training compared to using non-optimized seed selection.

Abstract: Physically-based volume rendering generates a lightfield. The locations of scattering modeled in physically-based rendering are used to assign depths for the lightfield. The previously assigned depths and previously rendered lightfield are used for lightfield rendering, which may be performed more rapidly than the physically-based volume rendering.

Abstract: Systems and methods are provided for generating a two-dimensional image for identification of medical imaging data. An image processor acquires the medical imaging data and determines a category of the medical imaging data. A machine-learnt network identifies as a function of the category, a plurality of settings of rendering parameters that highlight one or more features the medical imaging data. The image processor renders the two-dimensional identifier image from the medical imaging data using the plurality of settings of rendering parameters and stores the medical imaging data with the two-dimensional identifier image.

Abstract: An embodiment suggests a method for visualizing an image data set, in particular a medical image data set, wherein the visualized data set displays a three dimensional arrangement having at least a first object and a second object. The method includes assigning a first set of parameter to the first object; assigning a second set of parameters to the second object; dividing the medical image data set into a first sub-region and a second sub-region; and providing a visualisation of the three dimensional arrangement by a volume rendering method, in particular by a ray-casting method or a photorealistic volumetric path tracing, the first set of parameter being applied to the first sub-region for visualizing the first object and the second set of parameter being applied to the second sub-region for visualizing the second object.

Abstract: For three-dimensional rendering, a machine-learnt model is trained to generate representation vectors for rendered images formed with different rendering parameter settings. The distances between representation vectors of the images to a reference are used to select the rendered image and corresponding rendering parameters that provides a consistency with the reference. In an additional or different embodiment, optimized pseudo-random sequences are used for physically-based rendering. The random number generator seed is selected to improve the convergence speed of the renderer and to provide higher quality images, such as providing images more rapidly for training compared to using non-optimized seed selection.

Abstract: For interactive rendering in medical imaging, physically-based volume rendering of a volume of a patient may better assist physicians in diagnosis, prognosis, and/or planning. To provide for more rapid interaction, direct volume rendering is used during interaction. The rendering then transitions to physically-based rendering when there is no interaction. For smoothing the transition and/or preserving cohesive perceptual details, images from the different types of rendering may be blended in the transition and/or during interaction.

Abstract: For three-dimensional rendering, a machine-learnt model is trained to generate representation vectors for rendered images formed with different rendering parameter settings. The distances between representation vectors of the images to a reference are used to select the rendered image and corresponding rendering parameters that provides a consistency with the reference. In an additional or different embodiment, optimized pseudo-random sequences are used for physically-based rendering. The random number generator seed is selected to improve the convergence speed of the renderer and to provide higher quality images, such as providing images more rapidly for training compared to using non-optimized seed selection.

Abstract: A computer-implemented method of reducing 4D Digital Subtracted Angiography (DSA) reconstruction artifacts using a computational fluid dynamics (CFD) simulation includes a computer receiving first DSA time sequence data comprising a representation of a plurality of vessels and segmenting a vessel of interest from the first DSA time sequence data. The computer uses the CFD simulation to simulate fluid dynamics across the vessel of interest to yield a flow field and determines a plurality of simulated time activity curve parameters for each voxel inside the vessel of interest using the flow field. Then, the computer applies a reconstruction process to second DSA time sequence data to yield a DSA volume. This reconstruction process is constrained by the plurality of simulated time activity curve parameters for each voxel inside the vessel of interest.

Abstract: Physically-based volume rendering produces pixels. By assigning depths to the pixels, a 3D point cloud is generated. For more rapid rendering, such as due to user interaction, the 3D point cloud rendering is a proxy to physically-based volume rendering. The rendering parameters for the desired image may be determined using the proxy, and then physically-based volume rendering is performed to obtain the desired image.

Abstract: A computer-implemented method for performing window-leveling of volumetric ray tracing includes a computer system retrieving a quad-tree data structure organizing a plurality of light paths associated with a volume rendered using a transfer function and one or more window level values. The computer system also receives a user selection of at least one of a new transfer function or new window level values. The volume is rendered by the computer system with the new transfer function or the new window level values using the quad-tree data structure. Then, the computer system synthesizes an image by adding contributions from all of the plurality of light paths according to the quad-tree data structure.

Abstract: Systems and methods are provided for determining an analytic measure of a patient population. A knowledge database comprising structured patient data for a patient population is maintained. The structured patient data is generated by processing unstructured medical imaging data for the patient population using one or more machine learning algorithms. An analytic measure of the patient population is determined based on the structured patient data of the knowledge database. The analytic measure of the patient population is output.