A computer implemented method of generating at least one shape of a region of interest in a digital image is provided. The method includes obtaining, by an image processing engine, access to a digital tissue image of a biological sample; tiling, by the image processing engine, the digital tissue image into a collection of image patches; identifying, by the image processing engine, a set of target tissue patches from the collection of image patches as a function of pixel content within the collection of image patches; assigning, by the image processing engine, each target tissue patch of the set of target tissue patches an initial class probability score indicating a probability that the target tissue patch falls within a class of interest, the initial class probability score generated by a trained classifier executed on each target tissue patch; generating, by the image processing engine, a first set of tissue region seed patches by identifying target tissue patches having initial class probability scores that satisfy a first seed region criteria, the first set of tissue region seed patches comprising a subset of the set of target tissue patches; generating, by the image processing engine, a second set of tissue region seed patches by identifying target tissue patches having initial class probability scores that satisfy a second seed region criteria, the second set of tissue region seed patches comprising a subset of the set of target tissue patches; calculating, by the image processing engine, a region of interest score for each patch in the second set of tissue region seed patches as a function of initial class probability scores of neighboring patches of the second set of tissue region seed patches and a distance to patches within the first set of issue region seed patches; and generating, by the image processing engine, one or more region of interest shapes by grouping neighboring patches based on their region of interest scores.

A method for using an assembled three-dimensional image to construct a three-dimensional model for determining a path through a lumen network to a target. The three-dimensional model is automatically registered to an actual location of a probe by tracking and recording the positions of the probe and continually adjusting the registration between the model and a display of the probe position. The registration algorithm becomes dynamic (elastic) as the probe approaches smaller lumens in the periphery of the network where movement has a bigger impact on the registration between the model and the probe display.

Methods and systems for image processing are provided. Image data may be obtained. The image data may include a plurality of voxels corresponding to a first plurality of ribs of an object. A first plurality of seed points may be identified for the first plurality of ribs. The first plurality of identified seed points may be labelled to obtain labelled seed points. A connected domain of a target rib of the first plurality of ribs may be determined based on at least one rib segmentation algorithm. A labelled target rib may be obtained by labelling, based on a hit-or-miss operation, the connected domain of the target rib, wherein the hit-or-miss operation may be performed using the labelled seed points to hit the connected domain of the target rib.

The present disclosure provides methods and systems for image-based nerve fiber extraction. The methods may include obtaining an anatomical image of a subject and a diffusion image of the subject. The subject may include at least one region of interest (ROI) that relates to extraction of at least one target nerve fiber in the subject. The methods may further include determining, based on the anatomical image, the at least one ROI in the diffusion image; and extracting, from the diffusion image, at least one of the at least one target nerve fiber based on the at least one ROI.

An image-processing device includes: a gradient data generation unit which sequentially targets pixels, and on the basis of a luminance value, generates gradient data in which each pixel is represented using a first label of a value indicating a direction toward a pixel having a higher luminance value or a second label of a value indicating that the luminance value of the pixel is higher than all pixels located around the pixel; a plurality of region label generation units which set each pixel represented by the second label as a peak pixel, divide regions into regions including pixels belonging to the same peak pixel on the basis of peak coordinates indicating the positions of peak pixels, and generate region label data; and a peak pixel distribution unit which distributes each peak pixel to each of the region label generation units such that loads of computations are approximately equalized.

A computer implemented method of generating at least one shape of a region of interest in a digital image is provided. The method includes obtaining, by an image processing engine, access to a digital tissue image of a biological sample; tiling, by the image processing engine, the digital tissue image into a collection of image patches; identifying, by the image processing engine, a set of target tissue patches from the collection of image patches as a function of pixel content within the collection of image patches; assigning, by the image processing engine, each target tissue patch of the set of target tissue patches an initial class probability score indicating a probability that the target tissue patch falls within a class of interest, the initial class probability score generated by a trained classifier executed on each target tissue patch; generating, by the image processing engine, a first set of tissue region seed patches by identifying target tissue patches having initial class probability scores that satisfy a first seed region criteria, the first set of tissue region seed patches comprising a subset of the set of target tissue patches; generating, by the image processing engine, a second set of tissue region seed patches by identifying target tissue patches having initial class probability scores that satisfy a second seed region criteria, the second set of tissue region seed patches comprising a subset of the set of target tissue patches; calculating, by the image processing engine, a region of interest score for each patch in the second set of tissue region seed patches as a function of initial class probability scores of neighboring patches of the second set of tissue region seed patches and a distance to patches within the first set of issue region seed patches; and generating, by the image processing engine, one or more region of interest shapes by grouping neighboring patches based on their region of interest scores.

The present invention relates to the technical field of image processing, and in particular, to a patch net model for image representation and a construction method of the patch net model. The patch net model for image representation is of a forest-shaped structure consisting of a plurality of composite nodes and basic nodes, each composite node is a non-leaf node, and each basic node is a leaf node; the basic node includes a certain patch region of an image and a representative patch representing an apparent feature of the patch region; the composite node includes a certain patch region of the image and can be further decomposed into basic nodes and/or composite nodes; an edge exists between two nodes, which are located on the same layer of the forest-shaped structure and are spatially connected, and a relation matrix used for expressing the spatial relative position of the two nodes is arranged on the edge.

Disclosed are systems, devices, and methods for determining pleura boundaries of a lung, an exemplary method comprising acquiring image data from an imaging device, generating a set of two-dimensional (2D) slice images based on the acquired image data, determining, by a processor, a seed voxel in a first slice image from the set of 2D slice images, applying, by the processor, a region growing process to the first slice image from the set of 2D slice images starting with the seed voxel using a threshold value, generating, by the processor, a set of binarized 2D slice images based on the region grown from the seed voxel, filtering out, by the processor, connected components of the lung in each slice image of the set of binarized 2D slice images, and identifying, by the processor, the pleural boundaries of the lung based on the set of binarized 2D slice images.

In order to perform quantitative analysis on an object in an image, it is important to accurately identify the object, but when plural objects are in contact with each other, it is potential that a target portion cannot be accurately identified. An image is segmented into a foreground region and a background region, the foreground region being a region in which an object for which quantitative information is to be calculated is shown, and the background region being a region other than the foreground region. With respect to a first object and a second object in contact with each other in the image, a contact point between the first object and the second object is detected based on a region segmentation result output by a segmentation unit. The first object and the second object can be separated by connecting two boundary reference pixels including a first boundary reference pixel that is a pixel in a background region closest to the contact point, and a second boundary reference pixel that is a pixel in a background region in a direction opposite to the first boundary reference pixel across the contact point.

Novel tools and techniques are provided for implementing digital microscopy imaging using deep learning-based segmentation and/or implementing instance segmentation based on partial annotations. In various embodiments, a computing system might receive first and second images, the first image comprising a field of view of a biological sample, while the second image comprises labeling of objects of interest in the biological sample. The computing system might encode, using an encoder, the second image to generate third and fourth encoded images (different from each other) that comprise proximity scores or maps. The computing system might train an AI system to predict objects of interest based at least in part on the third and fourth encoded images. The computing system might generate (using regression) and decode (using a decoder) two or more images based on a new image of a biological sample to predict labeling of objects in the new image.