Neural network utilization with ultrasound technology

Patent metadata

Field	Value
Publication number	US-12236587-B2
Application number	US-202217671302-A
Country	US
Kind code	B2
Filing date	Feb 14, 2022
Priority date	Feb 14, 2022
Publication date	Feb 25, 2025
Grant date	Feb 25, 2025

Abstract

Examples herein include methods, systems, and computer program products for utilizing neural networks in ultrasound systems. The methods include processor(s) of a computing device identifying a neural network for implementation on the computing device to generate, based on ultrasound data, inferences and confidence levels for the inferences, the computing device being communicatively coupled via a computing network to an ultrasound machine configured to generate the ultrasound data. The processor(s) implements the neural network on the computing device, including configuring the neural network to generate an inference and a confidence level for at least one image of the images. The processor(s) obtains the ultrasound data including images from the ultrasound machine. The processor(s) determines, for the at least one image, an accuracy of the inference and the confidence level. The processor(s) automatically reconfigures the neural network to increase the accuracy based on the determining the accuracy.

First claim

What is claimed is:

1. A method implemented by a computing device, the method comprising:
identifying, by one or more processors of the computing device, a neural network for implementation on the computing device to generate, based on ultrasound data, inferences, the computing device being communicatively coupled via a computing network to an ultrasound machine configured to generate the ultrasound data;
obtaining, by the one or more processors, the ultrasound data including images from the ultrasound machine;
implementing, by the one or more processors, the neural network on the computing device, the implementing including configuring the neural network to generate an inference for at least one image of the images;
determining, for the at least one image, an accuracy of the inference; and
automatically reconfiguring, by the one or more processors, the neural network to increase the accuracy based on the determining the accuracy.

2. The method of claim 1 , wherein the inferences comprise confidence levels for the inferences.

3. The method of claim 1 , wherein the identifying the neural network includes:
identifying, by the one or more processors, an artificial intelligence (AI) instruction set for use in the implementing the neural network on the computing device, the identifying the AI instruction set including:
determining, by the one or more processors, if the computing device includes the AI instruction set, wherein the AI instruction set is selected from the group consisting of: a native AI instruction set and an AI instruction set from an application programming interface (API) supported by the computing device; and
based on determining that the computing device includes the AI instruction set, configuring, by the one or more processors, the neural network utilizing the AI instruction set.

4. The method of claim 3 , wherein the determining if the computing device includes the AI instruction set includes:
querying, by the one or more processors, at least one element of the computing device selected from the group consisting of: a configuration file, an operating system, a device identifier, and a resource manager.

5. The method of claim 1 , wherein the identifying the neural network includes:
identifying, by the one or more processors, an artificial intelligence (AI) instruction set for use in the implementing the neural network on the computing device, the identifying the AI instruction set including:
determining, by the one or more processors, if the computing device includes the AI instruction set, wherein the AI instruction set is selected from the group consisting of: a native AI instruction set and an AI instruction set from an application programming interface (API) supported by the computing device; and
based on determining that the computing device does not include the AI instruction set, enabling, by the one or more processors, a neural network of an ultrasound application executing on the computing device to generate the inference and for at least one image of the images.

6. The method of claim 1 , wherein the identifying the neural network includes:
querying, by the one or more processors, the computing device to determine hardware capabilities of the computing device;
based on the hardware capabilities of the computing device, identifying, by the one or more processors, from a central repository communicatively coupled to the computing device, a neural network model to run on the computing device; and
importing, from the central repository, the neural network of the neural network model.

7. The method of claim 6 , wherein the ultrasound machine includes the central repository.

8. The method of claim 7 , further comprising:
displaying, by the one or more processors, in a user interface of the computing device, for a given image of the images, a visual representation of the inference for the given image and the given image.

9. The method of claim 8 , further comprising:
obtaining, by the one or more processors, via the user interface of the computing device, a user input; and
instructing, by the one or more processors, displaying of the user input on an interface of the ultrasound machine.

10. The method of claim 8 , further comprising:
obtaining, by the one or more processors, via the user interface of the computing device, a user input; and
instructing, by the one or more processors, automatically adjusting one or more settings of the ultrasound machine, based on the user input.

11. The method of claim 1 , wherein the obtaining the ultrasound data includes:
obtaining, by the one or more processors, the ultrasound data including encoded data; and
decoding, by the one or more processors, the encoded data.

12. The method of claim 11 , wherein a method to encode the ultrasound data is selected from the group consisting of: encoding the ultrasound data in an image of decodable indicia, modulating a property of a carrier signal according to the ultrasound data, and modulating a logo to include the encoded data.

13. The method of claim 8 , wherein the communication link supports transfers of the ultrasound data at a frame rate greater than a rate at which the ultrasound machine generates the images, wherein the obtaining the ultrasound data includes obtaining indications of redundant frames in frames comprising the images, and wherein the generating the inference for at least one image of the images is based on the frames without the indications and not the redundant frames.

Assignees

• Fujifilm Sonosite Inc

Inventors

• Dhatt Davin
• Duffy Thomas
• Pely Adam
• White Christopher
• Lundberg Andrew
• Danset Paul
• Chamberlain Craig
• Mandavia Diku

Classifications

• G06N5/04

  Inference or reasoning models · CPC title

• G16H30/20 Primary

  for handling medical images, e.g. DICOM, HL7 or PACS · CPC title

• G06T2207/10132

  Ultrasound image · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title