Abstract: A computer-implemented method for generating a 2D or 3D object, including training an autoencoder on a first set of training data to identify a first set of latent variables and generate a first set of output data; training an hourglass predictor on a second set of training data, where the hourglass predictor encoder converts a set of related but different training input data to a second set of latent variables, which decode into a second set of output data of the same type as the first set of output data; and using the hourglass predictor to predict a 2D or 3D object of the same type as the first set of output data based on a 2D or 3D object of the same type as the second set of input data.

Abstract: A training system for an endoscopic procedure, the training system including an endoscope having an image sensor configured to generate images; an image processing device including a housing and a processing circuit in the housing, the processing circuit including a processor and memory, the memory having graphical user interface (GUI) logic and a trained single-pass neural network model, the processor being configured to process the neural network model and the GUI logic to present a GUI; and a physical model including cavities corresponding to interconnected human cavities. The GUI is configured to receive a user input corresponding to a training mode or an exam mode. In the training mode the GUI is configured to present navigation feedback, and in the exam mode the GUI is configured to not present navigation feedback.

Abstract: Disclosed is an image processing device for estimating a position of an endoscope in a model of the human airways using a first machine learning data architecture trained to determine a set of anatomic reference positions, said image processing device comprising a processing unit operationally connectable to an image capturing device of the endoscope, wherein the processing unit is configured to obtain a stream of recorded images; continuously analyse the recorded images of the stream of recorded images using the first machine learning data architecture to determine if an anatomic reference position of a subset of anatomic reference positions, from the set of anatomic reference positions, has been reached; and where it is determined that the anatomic reference position has been reached, update the endoscope position based on the anatomic reference position, and an endoscope system comprising an endoscope and an image processing device, a display unit comprising an image processing device, and a computer progr

Abstract: An image processing device including a housing and a processing circuit in the housing, the processing circuit including a processor and memory, the memory including a neural network model and proximity suppression logic, the neural network model comprising a single-pass neural network model trained with training images corresponding to an endoscopic procedure and defining anatomic references observable in the training images, wherein the neural network model is configured to process images, to detect the anatomic references in the images, and to output a set of anatomic references including identifiers and confidence values representing the likelihoods that the anatomic reference identifiers are correct, and wherein the proximity suppression logic is configured to change the confidence values of the anatomic references in the set of anatomic references based on a prior position of the endoscope to identify an anatomic reference indicative of the current position of the endoscope.

Abstract: A computer-implemented method for generating a corresponding 3D mesh representing a 3D object includes transforming an initial three-dimensional mesh into a planar mesh, wherein each vertex or edge of the planar mesh is a transformation of a vertex or edge from the initial three-dimensional mesh; and sampling the planar mesh to generate a plurality of samples such that each sample comprises a three-dimensional coordinate representing a point in a three-dimensional space derived and/or taken directly from the initial three-dimensional mesh, and a coordinate representing a position of the sample relative to other samples; and generating the sampled matrix based on the plurality of samples; and representing the sampled matrix as a corresponding 3D mesh.

Abstract: A computer-implemented method for generating a 2D or 3D object, including training an autoencoder on a first set of training data to identify a first set of latent variables and generate a first set of output data; training an hourglass predictor on a second set of training data, where the hourglass predictor encoder converts a set of related but different training input data to a second set of latent variables, which decode into a second set of output data of the same type as the first set of output data; and using the hourglass predictor to predict a 2D or 3D object of the same type as the first set of output data based on a 2D or 3D object of the same type as the second set of input data.

Abstract: A computer-implemented method for generating a corresponding 3D mesh representing a 3D object includes transforming an initial three-dimensional mesh into a planar mesh, wherein each vertex or edge of the planar mesh is a transformation of a vertex or edge from the initial three-dimensional mesh; and sampling the planar mesh to generate a plurality of samples such that each sample comprises a three-dimensional coordinate representing a point in a three-dimensional space derived and/or taken directly from the initial three-dimensional mesh, and a coordinate representing a position of the sample relative to other samples; and generating the sampled matrix based on the plurality of samples; and representing the sampled matrix as a corresponding 3D mesh.

Abstract: A computer-implemented method for generating a corresponding 3D mesh representing a 3D object, includes transforming an initial three-dimensional mesh into a planar mesh, wherein each vertex or edge of the planar mesh is a transformation of a vertex or edge from the initial three-dimensional mesh; and sampling the planar mesh to generate a plurality of samples such that each sample comprises a three-dimensional coordinate representing a point in a three-dimensional space derived and/or taken directly from the initial three-dimensional mesh, and a coordinate representing a position of the sample relative to other samples; and generating the sampled matrix based on the plurality of samples; and representing the sampled matrix as a corresponding 3D mesh.