# US 12,118,724 B2 · Interactive coronary labeling using interventional x-ray images and deep learning

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

- Number: US 12,118,724 B2
- Document ID: US12118724B2
- Status: Granted
- Kind: B2
- Application number: 18/508,319
- Filed: 2023-11-14
- Published: 2024-03-07
- Granted: 2024-10-15
- Earliest claimed: 2018-05-03
- Assignees: Koninklijke Philips N.V.
- Applicants: Koninklijke Philips N.V.
- Inventors: Roy Franciscus Petrus Van Pelt; Javier Olivan Bescos
- Web page: https://patentgrep.com/patent/US12118724B2

## CPC

- G16H 30/40: for processing medical images, e.g. editing
- G06T 7/0012: {Biomedical image inspection}
- G06T 11/60: Creating or editing images; Combining images with text
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G16H 50/50: for simulation or modelling of medical disorders
- G06T 2207/20081: Training; Learning
- G06T 2207/30101: Blood vessel; Artery; Vein; Vascular
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method for classifying a vasculature comprises training a training device with an initial model of a vasculature using diagnostic image data representing a geometry for a plurality of vessels of a vessel tree and including a respective vessel labeling for each vessel, providing at least one diagnostic image of a patient's vessel tree and identifying a variation between the vessel tree represented by the initial model and the patient's vessel tree. This variation is checked and labeled in order to improve the trained model. The process may be repeated iteratively until reaching an accurate patient-specific model of the vasculature.

## Claims

### Claim 1

A method of generating a model to classify a vasculature, the method comprising:
obtaining a trained model of a vasculature, wherein the trained model was trained from image data that represents a plurality of first vessel trees of a plurality of vasculatures and includes labeling of at least one vessel in each of the plurality of first vessel trees;
obtaining an image representing a second vessel tree;
identifying a deviation between a vessel geometry in the trained model and a vessel geometry in the second vessel tree;
providing a labeling of the deviation; and
adjusting, based on the deviation and the labeling of the deviation, the trained model to classify the vasculature.

### Claim 2 (depends on 1)

The method according to claim 1, further comprising:
displaying an indication of the deviation to a user; and
receiving user input providing the labeling of the deviation.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the adjusting of the trained model is iteratively repeating for each respective deviation of a plurality of deviations between a vessel geometry in the trained model and a vessel geometry of the respective deviation.

### Claim 4 (depends on 1)

The method according to claim 1, wherein the adjusting of the trained model comprises retraining the trained model based on the plurality of first vessel trees, the labeling of at least one vessel in each of the plurality of first vessel trees, the deviation between the vessel geometry in the trained model and the vessel geometry in the second vessel tree, and the labeling of the deviation.

### Claim 5 (depends on 1)

The method according to claim 1, further comprising identifying the vessel geometry of the second vessel tree in the image by:
segmenting the image,
identifying, based on the segmenting, centerline information for a plurality of vessels of the second vessel tree, and
extracting the centerline information for identification of the vessel geometry of the second vessel tree.

### Claim 6 (depends on 5)

The method according to claim 5, wherein the centerline information is obtained with the image representing the second vessel tree.

### Claim 7 (depends on 1)

The method according to claim 1, wherein the image data comprises between 10000 and 100 images.

### Claim 8 (depends on 1)

The method according to claim 1, wherein the trained model includes externally-caused variations.

### Claim 9

A system for generating a model to classify a vasculature, the classification system comprising:
a processor in communication with memory, the processor configured to:
obtain a trained model of the vasculature, wherein the trained model was trained from image data that represents a plurality of first vessel trees of a plurality of vasculatures and includes labeling of at least one vessel in each of the plurality of first vessel trees;
obtain an image representing a second vessel tree;
identify a deviation between a vessel geometry in the trained model and a vessel geometry in the second vessel tree;
provide a labeling of the deviation; and
adjust, based on the deviation and the labeling of the deviation, the trained model to classify the vasculature.

### Claim 10 (depends on 9)

The system according to claim 9, further comprising a display configured to:
display an indication of the deviation to a user, and
receive user input from the user indicating the labeling of the deviation.

### Claim 11 (depends on 9)

The system according to claim 9, wherein the processor is further configured to generate a medical representation comprising at least the vasculature, the labeling of at least one vessel in each of the plurality of first vessel trees, and the labeling of the deviation; and transmit the medical representation to a database.

### Claim 12 (depends on 11)

The system according to claim 11, wherein the database comprises an electronic medical record (EMR); and wherein the medical representation is generated according to a pre-determined format of the electronic medical record.

### Claim 13 (depends on 9)

The system according to claim 9, wherein the processor is further configured to:
segment the image,
identify, based on the segmenting, centerline information for a plurality of vessels of the second vessel tree, and
extract the centerline information for identification of the vessel geometry of the second vessel tree.

### Claim 14 (depends on 9)

The system according to claim 9, wherein the processor is further configured to iteratively repeat adjustment of the trained model for each respective deviation of a plurality of deviations between a vessel geometry in the trained model and a vessel geometry of the respective deviation.

### Claim 15 (depends on 9)

The system according to claim 9, wherein adjustment of the trained model comprises retraining the model based on the plurality of first vessel trees, the labeling of at least one vessel in each of the plurality of first vessel trees, the deviation between the vessel geometry in the trained model and the vessel geometry in the second vessel tree, and the labeling of the deviation.

### Claim 16

A non-transitory computer-readable medium having stored a computer program comprising instructions which, when executed by a processor, cause the processor to:
obtain a trained model of a vasculature, wherein the trained model was trained from image data that represents a plurality of first vessel trees of a plurality of vasculatures and includes labeling of at least one vessel in each of the plurality of first vessel trees;
obtain an image representing a second vessel tree;
identify a deviation between a vessel geometry in the trained model and a vessel geometry in the second vessel tree;
provide a labeling of the deviation; and
adjust, based on the deviation and the labeling of the deviation, the trained model to classify the vasculature.

### Claim 17 (depends on 16)

The non-transitory computer-readable medium of claim 16, wherein the instruction, when executed by the processor, further cause the processor to:
display an indication of the deviation to a user; and
receive user input providing the labeling of the deviation.

### Claim 18 (depends on 16)

The non-transitory computer-readable medium of claim 16, wherein the instruction, when executed by the processor, further cause the processor to:
segment the image,
identify, based on the segmenting, centerline information for a plurality of vessels of the second vessel tree, and
extract the centerline information for identification of the vessel geometry of the second vessel tree.

### Claim 19 (depends on 16)

The non-transitory computer-readable medium of claim 16, wherein the instruction, when executed by the processor, cause the processor to iteratively repeat adjustment of the trained model for each respective deviation of a plurality of deviations between a vessel geometry in the trained model and a vessel geometry of the respective deviation.

### Claim 20 (depends on 16)

The non-transitory computer-readable medium of claim 16, wherein adjustment of the trained model comprises retraining the model based on the plurality of first vessel trees, the labeling of at least one vessel in each of the plurality of first vessel trees, the deviation between the vessel geometry in the model and the vessel geometry in the second vessel tree, and the labeling of the deviation.