A system and method for image segmentation are provided. A three-dimensional image data set representative of a region including at least one airway may be acquired. The data set may include a plurality of voxels. A first-level seed within the region may be identified. A first-level airway within the region may be identified based on the first-level seed. A second-level airway may be identified within the region based on the first-level airway. The first-level airway and the second-level airway may be fused to form an airway tree.

A method for computing patient-specific hemodynamics. The method includes receiving three dimensional imaging data of a patent, extracting anatomical data from the three dimensional imaging data, calculating velocity and pressure fields corresponding to the extracted anatomical data, and calculating displacement and velocity of extracted solid particles corresponding to the anatomical data. The anatomical data comprises an anatomical boundary.

A computer-vision pipeline is organized as a closed loop of a sensor-processing phase, an image-processing phase, and an object-detection phase, each comprising a respective phase processor coupled to a master processor. The sensor-processing phase creates multiple exposure images, and derives multi-exposure multi-scale zonal illumination-distributions, to be processed independently in the image-processing phase. In a first implementation of the object-detection phase, extracted exposure-specific features are pooled prior to overall object detection. In a second implementation, exposure-specific objects, detected from the exposure-specific features, are fused to produce the sought objects of a scene under consideration. The two implementations enable detecting fine details of a scene under diverse illumination conditions. The master processor performs loss-function computations to derive updated training parameters of the processing phases. Several experiments applying a core method of operating the computer-vision pipelines, and variations thereof, ascertain performance gain under challenging illumination conditions.

An interactive image segmenting apparatus and method are provided. The image segmenting apparatus and corresponding method include a boundary detector, a condition generator, and a boundary modifier. The boundary detector is configured to detect a boundary from an image using an image segmentation process. The feedback receiver is configured to receive information about the detected boundary. The condition generator is configured to generate a constraint for the image segmentation process based on the information. The boundary modifier is configured to modify the detected boundary by applying the generated constraint to the image segmentation process.

A method and apparatus of detection, registration and quantification of an image. The method may include obtaining an image of a lithographically created structure, and applying a level set method to an object, representing the structure, of the image to create a mathematical representation of the structure. The method may include obtaining a first dataset representative of a reference image object of a structure at a nominal condition of a parameter, and obtaining second dataset representative of a template image object of the structure at a non-nominal condition of the parameter. The method may further include obtaining a deformation field representative of changes between the first dataset and the second dataset. The deformation field may be generated by transforming the second dataset to project the template image object onto the reference image object. A dependence relationship between the deformation field and change in the parameter may be obtained.

Image segmentation based on the combination of a deep learning network and a shape-guided deformable model is provided. In various embodiments, a time sequence of images is received. The sequence of images is provided to a convolutional network to obtain a sequence of preliminary segmentations. The sequence of preliminary segmentations labels a region of interest in each of the images of the sequence. A reference and auxiliary mask are generated from the sequence of preliminary segmentations. The reference mask corresponds to the region of interest. The auxiliary mask corresponds to areas outside the region of interest. A final segmentation corresponding to the region of interest is generated for each of the sequence of images by applying a deformable model to the composite mask with reference to the auxiliary mask.

Systems and methods for performing instance segmentation. A memory stores instructions for executing processes for performing instance segmentation and a processor configured to execute the instructions. The processes include: generating a learning objective that uses pair-wise relationships between pixels in an input image; sampling pixels in each object instance to determine whether the sampled pixels are within a same object instance; training a neural network using the learning objection, wherein the neural network is configured to make pixel-wise predictions and to assign a cluster index to each pixel of the input image, with each pixel cluster being an object instance; performing graph coloring to assign a color to each object instance, with adjacent object instances having different colors; performing connected component extraction to recover each object instance based on the graph coloring; and generating a rendered image having the assigned color applied to each object instance.

A computer-implemented method, an apparatus, and a system for estimating synthetic values of quantitative metrics are provided. They involve calculating new, more accurate boundaries using a classifier based on local intensity and spatial estimators, for the segmentation mask provided by a non-local means patch-based segmentation in a test image, and estimating for the pixels of interest at least one synthetic value of a quantitative metric using a given value of the quantitative metric assigned to the reference images and the boundaries. The method, apparatus, and system provide the advantage of generating synthetic values directly comparable against known values for given subjects or against predetermined scales for diagnostic or prognostic purposes. In the specific case of Alzheimer's disease, the invention stretches the predictive range up to two full decades, which constitutes a significant advance in the field of medical diagnostics.

A method for image segmentation includes acquiring a three-dimensional (3D) image that includes a plurality of two-dimensional (2D) images arranged in a spatial order. The method also includes determining a preliminary seed point in a first 2D image of the plurality of 2D images. The method further includes determining, based on the preliminary seed point, a final seed point in a second 2D image of the plurality of 2D images, and determining, based on the final seed point, a volume of interest (VOI) in the 3D image.

Systems and methods for performing instance segmentation. A memory stores instructions for executing processes for performing instance segmentation and a processor configured to execute the instructions. The processes include: generating a learning objective that uses pair-wise relationships between pixels in an input image; sampling pixels in each object instance to determine whether the sampled pixels are within a same object instance; training a neural network using the learning objection, wherein the neural network is configured to make pixel-wise predictions and to assign a cluster index to each pixel of the input image, with each pixel cluster being an object instance; performing graph coloring to assign a color to each object instance, with adjacent object instances having different colors; performing connected component extraction to recover each object instance based on the graph coloring; and generating a rendered image having the assigned color applied to each object instance.