Computerised tomography image processing

Patent metadata

Field	Value
Publication number	US-12198348-B2
Application number	US-202017637259-A
Country	US
Kind code	B2
Filing date	Aug 21, 2020
Priority date	Aug 23, 2019
Publication date	Jan 14, 2025
Grant date	Jan 14, 2025

Abstract

A method for training a machine learning image segmentation algorithm to segment structural features of a blood vessel in a computed tomography (CT) image is described herein. The method comprises receiving a labelled training set for the machine learning image segmentation algorithm. The labelled training set comprising a plurality of CT images, each CT image of the plurality of CT images showing a targeted region of a subject, the targeted region including at least one blood vessel. The labelled training set further comprises a corresponding plurality of segmentation masks, each segmentation mask labelling at least one structural feature of a blood vessel in a corresponding CT image of the plurality of CT images. The method further comprises training a machine learning image segmentation algorithm, using the plurality of NCT images and the corresponding plurality of segmentation masks, to learn features of the CT images that correspond to structural features of the blood vessels labelled in the segmentation masks, and output a trained image segmentation model. The method further comprises outputting the trained image segmentation model usable for segmenting structural features of a blood vessel in a CT image. Further methods are described herein for using the trained image segmentation model to segment structural features of blood vessels, and to establish the training set for training the machine learning image segmentation model. Computing apparatuses and computer readable media are also described herein.

First claim

The invention claimed is:

1. A method for training a machine learning image segmentation algorithm to segment structural features of a blood vessel in a computed tomography (CT) image, the method comprising:
receiving a labelled training set for the machine learning image segmentation algorithm, the labelled training set comprising:
a plurality of CT images, each CT image of the plurality of CT images showing a targeted region of a subject, the targeted region including at least one blood vessel; and
a corresponding plurality of segmentation masks, each segmentation mask labelling at least one structural feature of a blood vessel in a corresponding CT image of the plurality of CT images;

training a machine learning image segmentation algorithm, using the plurality of CT images and the corresponding plurality of segmentation masks, to learn features of the CT images that correspond to structural features of the blood vessels labelled by the segmentation masks, and output a trained image segmentation model; and
outputting the trained image segmentation model usable for segmenting structural features of a blood vessel in a CT image.

2. A method according to claim 1 , wherein the at least one blood vessel of the targeted region of the CT image includes the aorta, or
wherein the targeted region of the CT image includes an aortic aneurysm, or,
wherein the structural features of the blood vessel comprise one or more of: inner lumen, outer wall, intima/media, false lumen, calcification, thrombus, ulceration, atherosclerotic plaques, or
wherein the blood vessel comprises an artery, or
wherein the blood vessel comprises a vein.

3. A method according to claim 1 , wherein the computed tomography (CT) image includes a contrast CT image (CCT) or a non-contrast CT image (NCT).

4. A method according to claim 1 , wherein the method further comprises generating the labelled training set.

5. A method according to claim 1 , wherein the labelled training set has been established according to the method comprising:
receiving a plurality of CCT images, each CCT image showing a targeted region of a subject, the targeted region including at least one blood vessel; and
segmenting the plurality of CCT images to generate a corresponding plurality of segmentation masks, each segmentation mask labelling at least one structural feature of the at least one blood vessel in the corresponding CCT image;
wherein the labelled training set includes pairs of CCT images and the corresponding segmentation masks.

6. A method according to claim 1 , wherein the machine learning image segmentation algorithm comprises a neural network.

7. A method according to claim 1 , wherein each segmentation mask of the plurality of segmentation masks comprises a binary segmentation mask.

8. A non-transitory computer-readable medium having instructions stored thereon which, when executed by one or more processors, cause the one or more processors to implement a method for training a machine learning image segmentation algorithm according to claim 1 .

9. A computing apparatus for training a machine learning image segmentation algorithm to segment structural features of a blood vessel in a computed tomography (CT) image, the apparatus comprising:
one or more memory units; and
one or more processors configured to execute instructions stored in the one or more memory units to perform the method of claim 1 .

10. A non-transitory computer-readable medium having stored thereon computer-readable code representative of the trained image segmentation model of claim 1 .

11. A non-transitory computer-readable medium according to claim 10 , the computer readable medium further having instructions stored thereon which, when executed by one or more processors, cause the one or more processors to implement a method comprising:
providing a CT image to the trained image segmentation model, the trained image segmentation model trained to learn features of CT images that correspond to structural features of blood vessels; and
segmenting, using the trained image segmentation model, at least one structural feature of a blood vessel in the provided CT image.

12. A method for segmenting structural features of a blood vessel in a computed tomograph (CT) image, the method comprising; providing the CT image to a trained image segmentation model, wherein the trained image segmentation model has been trained according to the method comprising:
receiving a labelled training set for the machine learning image segmentation algorithm, the labelled training set comprising:
a plurality of CT images, each CT image of the plurality of CT images showing a targeted region of a subject, the targeted region including at least one blood vessel; and
a corresponding plurality of segmentation masks, each segmentation mask labelling at least one structural feature of a blood vessel in a corresponding CT image of the plurality of CT images;

training a machine learning image segmentation algorithm, using the plurality of CT images and the corresponding plurality of segmentation masks, to learn features of the CT images that correspond to structural features of the blood vessels labelled by the segmentation masks, and output a trained image segmentation model; and
outputting the trained image segmentation model usable for segmenting structural features of a blood vessel in the provided CT image; and
segmenting, using the trained image segmentation model, at least one structural feature of a blood vessel in the provided CT image.

13. A non-transitory computer readable medium having stored thereon segmentation data generated using a method according to claim 12 .

14. A computing apparatus for segmenting structural features of a blood vessel in a computed tomography (CT) image, the apparatus comprising:
one or more memory units; and
one or more processors configured to execute instructions stored in the one or more memory units to perform the method of claim 12 .

15. A method comprising:
sending a computed tomography (CT) image to a server, the CT image showing a targeted region of a subject including at least one blood vessel; and
receiving, from the server, at least one segmented structural feature of the at least one blood vessel by performing the method of claim 12 .

16. A computing apparatus configured to perform the method of claim 15 .

Assignees

• Univ Oxford Innovation Ltd

Inventors

• Lee Regent
• Chandrashekar Anirudh
• Grau Vicente
• Handa Ashok

Classifications

• G06T2207/30101

  Blood vessel; Artery; Vein; Vascular · CPC title

• G06T2207/30048

  Heart; Cardiac · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T2207/20021

  Dividing image into blocks, subimages or windows · CPC title