US 12,586,179 B1Granted

Physician-guided machine learning system for assessing medical images to facilitate locating of a historical twin

OwnerNeville Iraniinventor

Life of this patent

Estimating the term…

Loading patent…

Abstract

A computer-implemented method of evaluating a user image of a patient to enable identification of a historical twin of the patient. The method includes organizing a plurality of medical images in an archive and receiving from a medical professional each of: (i) a region of interest; (ii) a textual description; (iii) selections for binary criteria; and (iv) weights of weighable criteria. The method comprises using a natural language search to create a relevant set of medical images and creating an optimal set from the relevant set of medical images by discarding medical images from the relevant set based at least on the selections for binary criteria. The method includes image processing medical images in the optimal set using the weight of the features of the region of interest to create medical image results. The relevant set comprises less than ten percent of the medical images in the archive.

Claims20

Dependent claims

A computer-implemented method of evaluating a user image of a patient to enable identification of a historical twin of said patient, comprising:

(a) processing a plurality of medical images to determine an anatomic characterization of each of said plurality of medical images;
(b) organizing each of said plurality of medical images in an archive using said determined anatomic characterization and metadata;
(c) displaying said user image using a graphical user interface;
(d) receiving from a medical professional, in connection with said user image and via said graphical user interface, each of:
(i) a region of interest;
(ii) a medical-professional-entered textual description associated with said region of interest;
(iii) a natural language search parameter;
(iv) selections for binary criteria;
and (v) weights of weighable criteria, including a weight to be given to each of:
(v)(a) an anatomical location of said region of interest;
and (v)(b) features of said region of interest;
(e) using a natural language search configured using said natural language search parameter to create a relevant set of medical images from said archive;
(f) creating an optimal set of medical images from said relevant set of medical images by discarding medical images from said relevant set of medical images based on said selections for binary criteria, said weight of said anatomical location of said region of interest, and said medical-professional-entered textual description;
and
(g) image processing medical images in said optimal set using said weight of said features of said region of interest to create medical image results;
wherein, said relevant set comprises less than ten percent of said plurality of medical images in said archive.

13 more elements

A computer-implemented method of evaluating a user image of a patient to enable identification of a historical twin of said patient, comprising:

(a) organizing a plurality of medical images in an archive;
(b) displaying said user image using a graphical user interface;
(c) receiving from a medical professional, in connection with said user image and via said graphical user interface, each of:
(i) a plurality of regions of interest;
(ii) a medical-professional-entered textual description associated with at least one of said plurality of regions of interest;
(iii) selections for binary criteria;
(iv) priority information for prioritizing at least one of said plurality of regions of interest;
and (v) weights of weighable criteria, including a weight to be given to each of:
(v)(a) an anatomical location of at least one of said plurality of regions of interest;
and (v)(b) features of at least one of said plurality of regions of interest;
(d) using a natural language search to create a relevant set of medical images from said archive;
(e) creating an optimal set of medical images from said relevant set of medical images by discarding medical images from said relevant set of medical images based on said selections for binary criteria, said medical-professional-entered textual description, and said weight of said anatomical location of at least one of said plurality of regions of interest;
and
(f) image processing medical images in said optimal set using said weight of said features of at least one of said regions of interest to create medical image results;
wherein, said relevant set comprises less than ten percent of said plurality of medical images in said archive.

12 more elements

Dependencies are inferred from the claim text.

Facts

Application: 17/644,975Filed 17 Dec 2021 · continuation in part
Priority: 16 Mar 2021
Granted: 24 Mar 2026
Term ends: Not estimatedThe source records don’t support an estimate.

Classification28 codes

G06N 3/0464
Architecture, e.g. interconnection topology: convolutional networks [CNN, ConvNet]
G06T 7/0012
Inspection of images, e.g. flaw detection: biomedical image inspection
G06F 3/04845
For the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a…:…
G06F 16/532
Querying: query formulation, e.g. graphical querying