Automated ultrasound video interpretation of a body part with one or more convolutional neural networks

Patent metadata

Field	Value
Publication number	US-11446008-B2
Application number	US-201916540759-A
Country	US
Kind code	B2
Filing date	Aug 14, 2019
Priority date	Aug 17, 2018
Publication date	Sep 20, 2022
Grant date	Sep 20, 2022

Abstract

In an embodiment, an intelligent system includes an electronic circuit configured to execute a neural network, to detect at least one feature in an image of a body portion while executing the neural network, and to determine a respective position and a respective class of each of the detected at least one feature while executing the neural network. For example, such a system can execute a neural network to detect at least one feature in an image of a lung, to determine a respective position within the image of each detected feature, and to classify each of the detected features as one of the following: A-line, B-line, pleural line, consolidation, and pleural effusion.

First claim

The invention claimed is:

1. A method, comprising:
capturing an ultrasound video having a plurality of images of a body portion via an ultrasound transducer, wherein the images of the body portion includes an image of a lung;
executing a neural network by an electronic circuit coupled to the ultrasound transducer, the electronic circuit is configured to execute the neural network to:
detect at least one feature in each of the images;
determine a respective position and a respective class of each of the detected at least one feature, determining a respective position including determining a respective bounding box in which the detected at least one feature is disposed and determining a respective class including determining a respective probability that the detected at least one feature belongs to the respective class;
group the detected at least one feature in at least one cluster in response to the respective position and the respective class, each cluster corresponding to a respective actual feature; and
determine a grade for the respective actual feature in response to the cluster in which the respective actual feature is grouped; and
diagnosing a pathology of the lung in response to the respective determined class of each of the detected at least one feature;
wherein each detection of the at least one feature is represented by an output having the respective probability, a respective coordinate and a respective width of the respective bounding box.

2. The method of claim 1 wherein the images of the body portion include an image of a lung.

3. The method of claim 1 wherein determining a respective position of each of the detected at least one feature includes determining a size of the respective bounding box in which the detected at least one feature is disposed.

4. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes:
determining a respective probability that the detected at least one feature belongs to the respective class; and
determining that the detected at least one feature belongs to the respective class in response to the respective probability being greater than a threshold for the respective class.

5. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes:
determining probabilities that the detected at least one feature belongs to respective classes; and
determining that the detected at least one feature belongs to the one of the respective classes corresponding to the highest one of the probabilities.

6. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes an A-line.

7. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes a pleural line.

8. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes a pleural effusion.

9. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes a B-line.

10. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes merged B-lines.

11. The method of claim 1 wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes a consolidation.

12. The method of claim 1 , further comprising:
wherein determining a respective class of each of the detected at least one feature includes determining that at least one of the detected at least one feature includes a pleural effusion; and
determining a severity of the pleural effusion.

13. The method of claim 1 , further comprising:
wherein the images of the body portion includes an image of a lung; and
diagnosing a pathology of the lung in response to the respective determined class of each of the detected at least one feature.

14. The method of claim 1 , further comprising:
wherein the images of the body portion includes an image of a lung; and
diagnosing a pathology of the lung in response to the respective position and to the respective determined class of each of the detected at least one feature.

15. A system, comprising:
an ultrasound transducer configured to acquire an ultrasound video having a plurality of images of a body portion, wherein the images of the body portion includes an image of a lung; and
an electronic circuit coupled to the ultrasound transducer configured
to execute a neural network;
to detect at least one feature in each of the images while executing the neural network;
to determine a respective position and a respective class of each of the detected at least one feature while executing the neural network, determining a respective position including determining a respective bounding box in which the detected at least one feature is disposed and determining a respective class including determining a respective probability that the detected at least one feature belongs to the respective class;
to group the detected at least one feature in at least one cluster in response to the respective position and the respective class while executing the neural network, each cluster corresponding to a respective actual feature;
to determine a grade for the respective actual feature in response to the cluster in which the respective actual feature is grouped while executing the neural network; and
to diagnose a pathology of the lung in response to the respective determined class of each of the detected at least one feature;
wherein each detection of the at least one feature is represented by an output having the respective probability, a respective coordinate and a respective width of the respective bounding box.

16. The system of claim 15 wherein the neural network includes a convolutional neural network.

17. The system of claim 15 wherein the neural network includes a single-shot-detector convolutional neural network.

18. The system of claim 15 wherein the electronic circuit, while executing the neural network, is configured to detect at least one feature in an ultrasound image of a lung.

19. The system of claim 15 wherein the electronic circuit includes a control circuit.

20. The system of claim 15 wherein the image includes an M-mode image.

21. The system of claim 15 wherein the body portion includes a lung and the function is lung sliding.

22. A tangible, non-transitory computer-readable medium storing instructions that, when executed by a computing circuit, cause the computing circuit, or another circuit under control of the computing circuit, to execute a neural network:
to detect at least one feature in each of images of a body portion acquired by an ultrasound transducer, wherein the images of the body portion includes an image of a lung; and
to determine a respective position and a respective class of each of the detected at least one feature, determining a respective position including determining a respective bounding box in which the detect

Assignees

• Tokitae Llc

Inventors

• Mehanian Courosh
• Kulhare Sourabh
• Zheng Xinliang
• Wilson Benjamin K

Classifications

• A61B8/5207 Primary

  involving processing of raw data to produce diagnostic data, e.g. for generating an image · CPC title

• A61B8/08 Primary

  Clinical applications (A61B8/02, A61B8/04, A61B8/06 take precedence) · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06F18/2431

  Multiple classes · CPC title