# US 2021/0343015 A1 · Automated stereology for determining tissue characteristics

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0343015 A1
- Document ID: US20210343015A1
- Status: Published application
- Kind: A1
- Application number: 17/308,592
- Filed: 2021-05-05
- Published: 2021-11-04
- Earliest claimed: 2016-11-11
- Assignees: University of South Florida; Stereology Resource Center, Inc.
- Applicants: University of South Florida; Stereology Resource Center, Inc.
- Inventors: PETER RANDOLPH MOUTON; HADY AHMADY PHOULADY; DMITRY GOLDGOF; LAWRENCE O. HALL
- Web page: https://patentgrep.com/patent/US20210343015A1

## CPC

- G06T 7/0014: {using an image reference approach}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 3/096: Transfer learning
- G06N 20/00: Machine learning
- G06T 5/20: using local operators
- G06T 5/70: Denoising; Smoothing
- G06T 7/11: Region-based segmentation
- G06T 7/174: involving the use of two or more images
- G06T 7/97: {Determining parameters from multiple pictures (depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85)}
- G06T 17/205: {Re-meshing}
- G06V 10/82: using neural networks
- G06V 20/695: {Preprocessing, e.g. image segmentation}
- G06V 20/698: {Matching; Classification}
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/20152: Watershed segmentation
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

Systems and methods for automated stereology are provided. A method can include providing an imager for capturing a Z-stack of images of a three-dimensional (3D) object; constructing extended depth of field (EDF) images from the Z-stack of images; performing a segmentation method on the EDF images including estimating a Gaussian Mixture Model (GMM), performing morphological operations, performing watershed segmentation, constructing Voronoi diagrams and performing boundary smoothing; and determining one or more stereology parameters such as number of cells in a region.

## Claims

### Claim 1

A method for performing computerized stereology, comprising:
providing an imager for capturing a Z-stack of images of a three-dimensional (3D) object, the Z-stack of images being a sequence of images of the 3D object captured in increments having a step size along a z-axis of the 3D object;
constructing extended depth of field (EDF) images from the Z-stack of images;
performing a segmentation method on the EDF images including estimating a Gaussian Mixture Model (GMM), performing morphological operations, performing watershed segmentation, constructing Voronoi diagrams and performing boundary smoothing; and
determining one or more stereology parameters.

### Claim 2

A method for performing computerized stereology, comprising:
providing an imager for capturing a Z-stack of images of a three-dimensional (3D) object, the Z-stack of images being a sequence of images of the 3D object captured in increments having a step size along a z-axis of the 3D object, wherein the 3D object is a tissue sample;
constructing extended depth of field (EDF) images from the Z-stack of images;
performing a segmentation method on the EDF images including nucleus detection and segmentation, clump segmentation, cytoplasm segmentation, boundary approximation, course refinement, and fine refinement; and
determining one or more stereology parameters.

### Claim 3 (depends on 2)

The method for performing computerized stereology of claim 2, wherein the nucleus detection and segmentation includes blurring the EDF images using a 2-D adaptive noise-removal filter, and iteratively binarizing the EDF images starting with a low threshold to find seed points from different nuclei.

### Claim 4 (depends on 3)

The method for performing computerized stereology of claim 3, wherein the nucleus detection and segmentation includes removing small and concave regions after each binarization and adding remaining regions to a nucleus mask.

### Claim 5 (depends on 4)

The method for performing computerized stereology of claim 4, wherein the nucleus mask keeps nuclei segmented at each execution phase of the segmentation method.

### Claim 6 (depends on 2)

The method for performing computerized stereology of claim 2, wherein the clump segmentation includes learning a Gaussian Mixture Model (GMM) with two components on pixel intensities using an Expectation Maximization (EM) algorithm.

### Claim 7 (depends on 6)

The method for performing computerized stereology of claim 6, wherein a first Gaussian estimates a distribution of foreground pixel intensities and a second estimates background pixel intensities.

### Claim 8 (depends on 2)

The method for performing computerized stereology of claim 2, wherein cytoplasm segmentation includes approximating a cytoplasm boundary of a nucleus by assigning parts of the EDF images that have a focus measure similar to the nucleus and are relatively close.

### Claim 9 (depends on 2)

The method for performing computerized stereology of claim 2, wherein the course refinement includes applying a grid to the EDF images, and applying a limited level of concavity by defining reachability notation and removing grid squares that are not reachable from a nucleus centroid, followed by discretization.

### Claim 10 (depends on 2)

The method for performing computerized stereology of claim 2, wherein the fine refinement includes a pixel level iterative process and replacing each nucleus region's pixel intensity with a mean intensity of the nucleus region's outer boundary.

### Claim 11

A method for performing computerized stereology, comprising:
providing an imager for capturing a Z-stack of images of a three-dimensional (3D) object, the Z-stack of images being a sequence of images of the 3D object captured in increments having a step size along a z-axis of the 3D object;
constructing extended depth of field (EDF) images from the Z-stack of images;
performing clump segmentation on the EDF images by binarizing the EDF images using a threshold determined by estimating a Gaussian Mixture Model to pixel intensities;
preprocessing the EDF images by converting the EDF images into grayscale and opening by reconstruction followed by closing by reconstruction;
performing watershed segmentation on the EDF images, wherein regional minimas are extracted as foreground markers and boundaries between regions are used as background markers, and the watershed segmentation is applied using the background and foreground makers that overlap with clumps;
constructing Voronoi diagrams and smoothing, including constructing a Voronoi map using centers of foreground regions and refining region boundaries using a Savitzy-Golay filter; and
determining one or more stereology parameters.

### Claim 12 (depends on 11)

The method for performing computerized stereology of claim 11, wherein the clump segmentation includes segmenting clumps of regions in the EDF images by a GMM with two components estimated based on pixel intensities using an Expectation Maximization Algorithm.

### Claim 13 (depends on 11)

The method for performing computerized stereology of claim 11, wherein the preprocessing includes smoothing the EDF images and removing small dark or bright regions.

### Claim 14 (depends on 12)

The method for performing computerized stereology of claim 12, wherein the preprocessing includes connecting relatively close regions and removing small region minimas.

### Claim 15 (depends on 11)

The method for performing computerized stereology of claim 11, wherein the foreground and background markers are region minimas extracted from preprocessed EDF images.

### Claim 16 (depends on 11)

The method for performing computerized stereology of claim 11, wherein the watershed segmentation expands original regional minimas to give a better approximation of neuron boundaries.

### Claim 17 (depends on 11)

The method for performing computerized stereology of claim 11, wherein the constructing Voronoi diagrams and smoothing includes not splitting a region if the region's size is less than a maximum threshold and solidity of the region obtained by the refined boundary of an original region is greater than an average solidity of all regions.

### Claim 18 (depends on 17)

The method for performing computerized stereology of claim 17, wherein the constructing Voronoi diagrams and smoothing includes not splitting a region if the region's size is less than a maximum threshold and solidity of the region obtained by the refined boundary of an original region is greater than an average solidity of all regions.

### Claim 19 (depends on 18)

The method for performing computerized stereology of claim 18, wherein in determining a number of cells, segmented regions are removed that do not overlap with a region of interest or overlap exclusion lines of a disector frame.

### Claim 20 (depends on 19)

The method for performing computerized stereology of claim 19, wherein a total number of cells (N) is determined according to the following equation:
Total N=[ΣQ−]·F1·F2·F3
wherein F1 is the reciprocal of the section sampling fraction (ssf); F2 is the reciprocal of the area sampling fraction (asf); and F3 is the reciprocal of the thickness sampling fraction (tsf).

### Claim at position 21

21-29. (canceled)
