Identification of inflammation in tissue images

Patent metadata

Field	Value
Publication number	US-10121245-B2
Application number	US-201615264836-A
Country	US
Kind code	B2
Filing date	Sep 14, 2016
Priority date	Sep 14, 2015
Publication date	Nov 6, 2018
Grant date	Nov 6, 2018

Abstract

Systems and methods are provided for identifying markers for inflammation in a tissue image. The tissue image is captured as an image of a histology slide. Subcellular structures in the tissue image are segmented via a first automated process to identify at least one variety of immune cells within the image. Glands and vilii are identified within the tissue image via a second automated process. Neutrophils are identified within the tissue image via a third automated process. An output representing the identified glands, villi, neutrophils, and other immune cells is provided to a human operator.

First claim

What is claimed is:

1. A method of identifying markers for inflammation in a tissue image comprising:
capturing the tissue image as an image of a histology slide;
segmenting subcellular structures in the tissue image, via a first automated process, to identify at least one variety of immune cells within the tissue image;
identifying glands and villi within the tissue image via a second automated process;
identifying neutrophils within the tissue image via a third automated process; and
providing an output representing the identified at least one variety of immune cells, the identified glands, the identified villi, and the identified neutrophils to a human operator.

2. The method of claim 1 , wherein segmenting subcellular structures in the tissue image via the first automated process comprises:
constructing a layered graph model, comprising a plurality of vertices and a plurality of edges according to at least one constraint;
determining respective weights for each of the plurality of vertices and the plurality of edges according to at least one photometric prior; and
determining a set of boundaries, represented by a path having a lowest total energy in the layered graph model, for a nucleus and a cytoplasm of a cell represented by the layered graph model.

3. The method of claim 2 , further comprising identifying the cell as one of a plasma cell, a lymphocyte, and an eosinophil from the determined set of boundaries.

4. The method of claim 1 , wherein identifying glands and villi within the tissue image via the second automated process comprises:
generating a superpixel segmentation of the image comprising a plurality of superpixels;
classifying each of the plurality of superpixels as one of epithelium, lumen, and extracellular material;
generating an initial pseudo-probability map from the classified plurality of superpixels;
generating a plurality of candidate objects in the tissue image according to the generated initial pseudo-probability map; and
classifying each candidate object by an associated pattern recognition classifier.

5. The method of claim 1 , wherein identifying neutrophils within the tissue image via the third automated process comprises:
generating a segmentation of the tissue image such that the lobes of each cell of a plurality of cells are grouped into one segment, which contains no lobes from other cells, to provide a plurality of segments;
classifying each of the plurality of segments to identify a plurality of high confidence examples of neutrophils and non-neutrophil cells;
applying a clustering process, utilizing a Voronoi diagram of clusters model, to the classified segments to locate high-confidence and low-confidence examples of neutrophils and non-neutrophil cells;
generating a plurality of classifiers, each representing a Voronoi diagram cell; and
classifying each ambiguous segment of the plurality of segments by at least one of the plurality of classifiers as a neutrophil or a non-neutrophil.

Assignees

• Univ Notre Dame Du Lac
• Univ California

Inventors

• Chen Danny Ziyi
• Wang Jiazhuo
• Mackenzie John Dewolfe
• Ramachandran Rageshree

Classifications

• G06V10/46

  Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features (colour feature extraction G06V10/56) · CPC title

• G06T2207/10056

  Microscopic image · CPC title

• G06K9/00147

  Physics · mapped topic

• G06T7/162

  involving graph-based methods · CPC title

• G06K9/48

  Physics · mapped topic