Abstract: Presently disclosed are systems and methods for extrapolating a predictive two-dimensional (2D) segmentation to indicate a predicted region of interest (ROI) in a 2D medical image frame rendered from a set of three-dimensional (3D) medical scan data, wherein the predictive segmentation is based on a user-defined 2D segmentation that annotates the ROI in another image frame rendered from the 3D medical scan data. Present embodiments can utilize a user-defined or user-modified 2D segmentation of a frame to extrapolate predicted 2D segmentations for all of the other frames within a set of 3D medical scan data. This enables present embodiments to effectively annotate the ROI in a series of 2D medical image frames with minimal input from a clinician, reducing the amount of time the clinician spends analyzing the 3D medical scan data.

Abstract: Systems/techniques that facilitate dynamic user-interface comparison between machine learning output and training data are provided. In various embodiments, a system can access a data candidate on which a machine learning classifier is to be executed. In various aspects, the system can generate, via execution of the machine learning classifier, a predicted classification based on the data candidate. In various instances, the system can render, on an electronic display, a confidence graph that visually depicts how well or how poorly the data candidate and the predicted classification fit into an annotated training dataset on which the machine learning classifier has been trained.

Abstract: Techniques are described that facilitate interactive segmentation of medical images with low latency in a cloud-based medical imaging application. According to an embodiment, a system is provided that comprises a memory that stores computer executable components, and a processor that executes the computer executable components stored in the memory. The computer executable components comprise a reception component that receives a compressed representation of a three-dimensional (3D) medical image generated from the 3D medical image by an encoder network of a deep neural network model comprising an encoder network and a decoder network, an interaction component that receives user input relative to a displayed portion of the 3D medical image indicating an object included in the displayed portion for segmentation, and a segmentation component that generates a segmentation mask for the object based on the user input using the compressed representation and the decoder network.

Abstract: Methods and apparatus are provided for processing data representing three-dimensional points organized in a data structure wherein each point has multiple components, the data is organized in a respective layer per component, each layer is segmented in cells of a two-dimensional grid, the cells are arranged such that the components of a given point are contained in corresponding cells of multiple layers, the cells are grouped in patches by layer, and the patches are arranged such that the components of an array of points is represented by corresponding patches of multiple layers. At least one first criterion and at least one second criterion are obtained. Data are retrieved from cells of patches meeting the at least one first criterion and from layers meeting the at least one second criterion. The retrieved data are processed to obtain a derivative data set.

Abstract: Methods and apparatus are provided for processing data representing three-dimensional points organized in a data structure wherein each point has multiple components, the data is organized in a respective layer per component, each layer is segmented in cells of a two-dimensional grid, the cells are arranged such that the components of a given point are contained in corresponding cells of multiple layers, the cells are grouped in patches by layer, and the patches are arranged such that the components of an array of points is represented by corresponding patches of multiple layers. At least one first criterion and at least one second criterion are obtained. Data are retrieved from cells of patches meeting the at least one first criterion and from layers meeting the at least one second criterion. The retrieved data are processed to obtain a derivative data set.