A method and system of automated segmentation of an anatomical object through learned examples include: receiving, by a processing device, an image of the anatomical object; determining a sparse representation of a shape of the anatomical object by iteratively evolving a segmenting surface as a combination of a level set segmentation and a linear combination of training shapes; and outputting, to an output device, the sparse representation of the shape of the anatomical object as the segmentation of the anatomical object.

A method for implementing a dynamic three-dimensional lung map view for navigating a probe inside a patient's lungs includes loading a navigation plan into a navigation system, the navigation plan including a planned pathway shown in a 3D model generated from a plurality of CT images, inserting the probe into a patient's airways, registering a sensed location of the probe with the planned pathway, selecting a target in the navigation plan, presenting a view of the 3D model showing the planned pathway and indicating the sensed location of the probe, navigating the probe through the airways of the patient's lungs toward the target, iteratively adjusting the presented view of the 3D model showing the planned pathway based on the sensed location of the probe, and updating the presented view by removing at least a part of an object forming part of the 3D model.

Medical imaging device- and display-invariant segmentation and measurement is provided. In various embodiments, a plurality of medical images is read from a data store. Metadata of each of the plurality of medical images is read. The metadata identifies an image acquisition device associated with each of the plurality of medical images. Based on the plurality of medical images and the metadata of each of the plurality of images, a learning system is trained to determine one or more image correction parameters. The one or more image correction parameters optimize segmentation of the plurality of medical images.

A method is provided to perform local de-noising of an image. The method includes obtaining a region of interest and a region of noise within a scan. The method also includes determining, for a first image based on the region of interest and a second image based on the region of noise, sample blocks and atoms for each image, where each atom contributes to a weighted sum that approximates a sample block in the image. The method also includes determining a measure of similarity of each atom from the first image with atoms from the second image and removing an atom from the first image if the measure of similarity exceeds a predetermined threshold value. The method also includes reconstructing a de-noised image based on atoms remaining in the first image after removing the atom from the first image, and presenting the de-noised image on a display device.

Neural network classification may be performed by inputting a training data set into each of a plurality of first neural networks, the training data set including a plurality of samples, obtaining a plurality of output value sets from the plurality of first neural networks, each output value set including a plurality of output values corresponding to one of the plurality of samples, each output value being output from a corresponding first neural network in response to the inputting of one of the samples of the training data set, inputting the plurality of output value sets into a second neural network, and training the second neural network to output an expected result corresponding to each sample in response to the inputting of a corresponding output value set.

A method of segmenting an MR organ volume includes performing regional mapping on the MR organ volume using a spatial prior probability map of a location of the organ to create a regionally mapped MR organ volume, and performing boundary refinement on the regionally mapped MR organ volume using a level set framework that employs the spatial prior probability map and a propagated shape constraint to generate a segmented MR organ volume.

The invention provides a method and device for creating a model for classifying a data point in imaging data representing measured intensities, the method comprising: training a model using a first labelled set of imaging data points; determining at least one first image part in the first labelled set which the model incorrectly classifies; generating second image parts similar to at least one image part; further training the model using the second image parts. Preferably the imaging data points and the second image parts comprise 3D data points.

Embodiments classify lung nodules by accessing a 3D radiological image of a region of tissue, the 3D image including a plurality of voxels and slices, a slice having a thickness; segmenting the nodule represented in the 3D image across contiguous slices, the nodule having a 3D volume and 3D interface, where the 3D interface includes an interface voxel; partitioning the 3D interface into a plurality of nested shells, a nested shell including a plurality of 2D slices, a 2D slice including a boundary pixel; extracting a set of intra-perinodular textural transition (Ipris) features from the 2D slices based on a normal of a boundary pixel of the 2D slices; providing the Ipris features to a machine learning classifier which computes a probability that the nodule is malignant, based, at least in part, on the set of Ipris features; and generating a classification of the nodule based on the probability.

A method includes receiving a time series of slice images of medical imaging. The images have a region of interest located at a lung lesion. The method also includes tracking over at least one subset of slice images in a time series of slice images variations over time of at least one image parameter at the set of points in the region of interest. Classifier processing is applied to set of signals indicative of tracked time variations of the at least one image parameter at respective points in the set of points. A classification signal is indicative of the tracked time variations of the at least one image parameter reaching or failing to reach at least one classification threshold.

Provided is a method and apparatus for reconstructing a medical image. The apparatus for reconstructing a medical image generates at least one base image by reducing a dimensionality of a three-dimensional (3D) medical image, generates at least one segmented image by reducing a dimensionality of a 3D image of a region of a tissue segmented from the 3D medical image or a 3D image of a region excluding the tissue from the 3D medical image, and trains, by using training data including the at least one base image and the at least one segmented image, an artificial intelligence (AI) model that separates at least one tissue from a medical image showing a plurality of tissues overlapping one another on the same plane.