Granted Patent B2

US 12,496,038 · App. 18/524,838 · Granted Dec 16, 2025

Inventors: Ross Upton (Oxford, GB); Paul Leeson (Oxford, GB)

Assignee: OXFORD UNIVERSITY INNOVATION LTD.

A61B8/0883A61B8/5223G06T7/0012A61B5/0044G06T2207/10132G06T2207/30048G16H30/40

Abstract

Claims (27)

1 . A method for diagnosing a heart condition comprising:

analyzing, by a processor, a set of images, wherein each of the images has a diagnostic outcome associated with it;

calculating, by the processor and based at least in part on the set of images, a one or more values of at least one parameter for each of the images using at least one of a principal transformation or a shear transformation;

analyzing, by the processor, the one or more values and the diagnostic outcomes to determine a relationship or correlation between the one or more values and the diagnostic outcomes;

using machine learning to develop a decision tree for generating the diagnostic outcomes from the values of the parameters, wherein the decision tree comprises a plurality of decision points and each decision point defines a reference or threshold value of at least one of the parameters; and

outputting a diagnosis of a heart condition based on the decision tree,

wherein each of the images is obtained under stress echocardiography conditions.

2 . The method according to claim 1 , further comprising:

storing images as respective data sets;

processing the respective data sets to determine locations of pairs of points of a heart; and

recording said locations of said pairs of points.

3 . The method according to claim 1 , wherein the decision tree is a C4.5 or a J48 pruned decision tree.

4 . The method according to claim 1 , wherein the decision tree comprises determining the value of at least one parameter of the heart for each of the images, wherein each of the images is obtained under stress echocardiography conditions.

5 . A system comprising:

a display; and

a processing unit configured to:

receive a set of images, wherein each of the images has an associated diagnostic outcome;

determine, based at least in part on the set of images, one or more values of at least one parameter for each of the images using at least one of a principal transformation or a shear transformation;

determine a relationship between the one or more values and the diagnostic outcomes;

generate, using machine learning, a decision tree for identifying a diagnosis of a heart condition from the values of the parameters, wherein the decision tree comprises a plurality of decision points and each decision point defines a reference or threshold value of at least one of the parameters; and

output, on the display, the diagnosis of the heart condition.

6 . The system of claim 5 , wherein the processing unit is further configured to:

store images as respective data sets;

process the respective data sets to determine locations of pairs of points of a heart; and

record said locations of said pairs of points.

7 . The system of claim 5 , wherein the decision tree is a C4.5 or a J48 pruned decision tree.

8 . The system of claim 5 , wherein each of the images is obtained under stress echocardiography conditions.

Assignments (1)

Priority Claims (1)

Continuity (3)

References Cited (80)