System for analysis of microscopic data using graphs

Patent metadata

Field	Value
Publication number	US-2021166387-A1
Application number	US-201917267821-A
Country	US
Kind code	A1
Filing date	Aug 13, 2019
Priority date	Aug 15, 2018
Publication date	Jun 3, 2021
Grant date	—

Abstract

Disclosed is a system for analysis of microscopic image data acquired from biological cells. The system includes a data processing system which is configured to read the image data and determine a plurality of vertices, wherein each of the vertices represents a location of an entity of interest within a region of interest of the image data. The data processing system generates a plurality of graphs, wherein for each of the graphs, the generation of the respective graph includes generating a plurality of edges, wherein each of the edges has two of the plurality of vertices associated therewith. For each of the graphs one or more vertex sets are identified, each of which consisting of one or more of the plurality of vertices. The data processing system further determines, for each of the graphs, a number of the identified vertex sets.

First claim

1 . A system for analysis of microscopic image data acquired from biological cells, the system comprising a data processing system which is configured to:
read the image data; determine a plurality of vertices, wherein each of the vertices represents a location of an entity of interest within a region of interest of the image data; generate a plurality of graphs, wherein for each of the graphs, the generation of the respective graph comprises generating a plurality of edges, wherein each of the edges has two of the plurality of vertices associated therewith; wherein for each of the graphs, the generating of the respective graph comprises: determining for a pair of the vertices whether to generate an edge associated with the pair depending on whether a distance between the vertices of the pair according to a predefined metric is smaller than a threshold which is predefined for the respective graph; wherein the predefined thresholds of the graphs are different from each other; wherein each of the graphs comprises the plurality of edges generated for the respective graph and the plurality of vertices; identify, for each of the graphs and depending on the edges generated for the respective graph, one or more vertex sets, each of which comprising one or more of the plurality of vertices; and to generate, for each of the graphs, a number of the identified vertex sets; characterized in that the data processing system is further configured to determine a difference between two of the determined numbers of identified vertex sets of different graphs; and to classify and/or to rate at least one tissue portion contained in the region of interest, a cell contained in the region of interest, a group of cells contained in the region of interest and/or the region of interest depending on the determined difference.

2 . The system of claim 1 , wherein the data processing system is further configured to determine a sum of the determined numbers of identified vertex sets over all or over a portion of the generated graphs.

3 . The system of claim 2 , wherein the data processing system is further configured to classify and/or to rate at least one tissue portion contained in the region of interest, a cell contained in the region of interest, a group of cells contained in the region of interest and/or the region of interest depending on the determined sum.

4 - 5 . (canceled)

6 . The system of claim 1 ,
wherein the generation of the graphs comprises: generating a Voronoi diagram depending on the vertices; wherein the Voronoi diagram comprises exactly one Voronoi region for each of the vertices; and wherein each of the graphs is generated based on the Voronoi diagram.

7 . The system of claim 6 , wherein each of the graphs is constructed from a respective radius-bounded Voronoi diagram, in which each of the Voronoi regions is spatially bounded by a same maximum Radius (R3);
wherein each of the graphs has a different maximum Radius (R3).

8 . The system of claim 1 , wherein the data processing system is further configured to classify and/or to rate at least one tissue portion contained in the region of interest, a cell contained in the region of interest, a group of cells contained in the region of interest and/or the region of interest depending on the determined numbers of generated vertex sets.

9 . (canceled)

10 . The system of claim 1 , wherein for each of the graphs, the generation of the plurality of edges comprises: determining for each pair of the vertices, whether or not to generate an edge.

11 . The system of claim 1 wherein for each of the graphs:
a) each of the identified vertex sets consists of one vertex or more vertices so that each pair thereof is connected by one or more of the edges of the graph; and
b) the identified vertex sets are mutually unconnected by the edges of the graph.

12 . The system of claim 1 , wherein the system comprises an image acquisition unit which is configured to:
receive a sample, which comprises the cells; and to image the cells.

13 . A method of analyzing microscopic data acquired from biological cells, wherein the analysis is performed using a data processing system, the method comprising:
reading using the data processing system, the image data; determining using the data processing system, a plurality of vertices, wherein each of the vertices represents a location of an entity of interest within a region of interest of the image data; generating, using the data processing system, a plurality of graphs, wherein for each of the graphs, the generation of the respective graph comprises generating a plurality of edges, wherein each of the edges has two of the plurality of vertices associated therewith; wherein for each of the graphs, the generating of the respective graph comprises: determining for a pair of the vertices whether to generate an edge associated with the pair depending on whether a distance between the vertices of the pair according to a predefined metric is smaller than a threshold (R 1 , R 2 ) which is predefined for the respective graph; wherein the predefined thresholds (R 1 , R 2 ) of the graphs are different from each other; wherein each of the graphs comprises the plurality of edges generated for the respective graph and the plurality of vertices; identifying, using the data processing system, for each of the graphs and depending on the edges generated for the respective graph, one or more vertex sets, each of which comprising one or more of the plurality of vertices; and generating, using the data processing system, for each of the graphs, a number of the identified vertex sets; characterized in that the method further comprises determining, using the data processing system, a difference between two of the determined numbers of identified vertex sets of different graphs; and classifying and/or rating, using the data processing system, at least one tissue portion contained in the region of interest, a cell contained in the region of interest, a group of cells contained in the region of interest and/or the region of interest depending on the determined difference.

14 . A program element for analyzing microscopic data acquired from biological cells, wherein the analysis is performed using a data processing system, wherein the program element, when being executed by a processor of the data processing system, is adapted to carry out:
reading the image data; determining a plurality of vertices, wherein each of the vertices represents a location of an entity of interest within a region of interest of the image data; generating, using the data processing system, a plurality of graphs, wherein for each of the graphs, the generation of the respective graph comprises generating a plurality of edges, wherein each of the edges has two of the plurality of vertices associated therewith; wherein for each of the graphs, the generating of the respective graph comprises: determining for a pair of the vertices whether to generate an edge associated with the pair depending on whether a distance between the vertices of the pair according to a predefined metric is smaller than a threshold (R 1 , R 2 ) which is predefined for the respective graph; wherein the predefined thresholds (R 1 , R 2 ) of the graphs are different from each other; wherein each of the graphs comprises the plurality of edges generated for the respective graph and the plurality of vertices; identifying, for each of the graphs and depending on the edges generated for the respective graph, one or more vertex sets, each of which comprising one or more of the plurality of vertices; and generating, for each of the graphs

Assignees

• Koninklijke Philips Nv

Inventors

• Chatterjea Supriyo
• Korst Johannes Henricus Maria
• Van Leeuwen Marinus Bastiaan
• Wimberger-Friedl Reinhold

Classifications

• G06F18/2414

  Smoothing the distance, e.g. radial basis function networks [RBFN] · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title

• G06T2207/30024

  Cell structures in vitro; Tissue sections in vitro · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G06T2207/20072

  Graph-based image processing · CPC title