Capsule endoscope for determining lesion area and receiving device

Patent metadata

Field	Value
Publication number	US-11715201-B2
Application number	US-202117339574-A
Country	US
Kind code	B2
Filing date	Jun 4, 2021
Priority date	Jun 14, 2017
Publication date	Aug 1, 2023
Grant date	Aug 1, 2023

Abstract

Provided is a capsule endoscope. The capsule endoscope includes: an imaging device configured to perform imaging on a digestive tract in vivo to generate an image; an artificial neural network configured to determine whether there is a lesion area in the image; and a transmitter configured to transmit the image based on a determination result of the artificial neural network.

First claim

What is claimed is:

1. A capsule endoscope comprising:
an imaging device configured to perform imaging on a digestive tract in a living body to generate an image;
an artificial neural network configured to determine the image is a valid image when a lesion area is present in the image; and
a transmitter configured to transmit only the valid image to a receiver.

2. The capsule endoscope of claim 1 , wherein the artificial neural network is based on a convolution neural network (CNN).

3. The capsule endoscope of claim 2 , wherein the artificial neural network stores a kernel matrix for determining pixels in the image and a weight indicating a degree of coupling of neurons in the CNN.

4. The capsule endoscope of claim 3 , wherein the kernel matrix and the weight are previously learned data.

5. The capsule endoscope of claim 4 , wherein the kernel matrix and the weight are updated by the artificial neural network.

6. The capsule endoscope of claim 1 , wherein the artificial neural network activates the transmitter if the lesion area is present in the image and generates a control signal for deactivating the transmitter if the lesion area is not present in the image.

7. The capsule endoscope of claim 6 , further comprising:
a power supply circuit configured to supply power to the imaging device, the artificial neural network, and the transmitter; and
a switch configured to connect the power supply circuit and the transmitter.

8. The capsule endoscope of claim 7 , wherein the switch connects the power supply circuit and the transmitter in accordance with the control signal.

9. The capsule endoscope of claim 1 , wherein the artificial neural network is based on spiking neural network (SNN).

10. The capsule endoscope of claim 9 , wherein the artificial neural network stores a weight indicating a degree of coupling of neurons in the SNN.

11. A method of a capsule endoscope, the method comprising:
performing imaging on a digestive tract in a living body to generate an image;
determining the image is a valid image when a lesion area is present in the image, by an artificial neural network of the capsule endoscope; and
transmitting only the valid image to a receiver.

12. The method of claim 11 , wherein determining the image is the valid image is based on a convolution neural network (CNN).

13. The method of claim 12 , wherein further comprising:
storing a kernel matrix for determining pixels in the image and a weight indicating a degree of coupling of neurons in the CNN.

14. The method of claim 13 , wherein the kernel matrix and the weight are previously learned data.

15. The method of claim 14 , wherein the kernel matrix and the weight are updated by an artificial neural network of the capsule endoscope.

16. The method of claim 11 , wherein the transmitting the valid image to a receiver comprises:
activating a transmitter of the capsule endoscope if the lesion area is present in the image; and deactivating the transmitter of the capsule endoscope if the lesion area is not present in the image.

17. The method of claim 11 , wherein the transmitting the valid image to a receiver comprises:
supplying a power to a transmitter of the capsule endoscope if the lesion area is present in the image.

18. The method of claim 11 , wherein determining the image is the valid image is based on a spiking neural network (SNN).

19. The method of claim 18 , wherein further comprising:
storing a weight indicating a degree of coupling of neurons in the SNN.

Assignees

• Electronics & Telecommunications Res Inst

Inventors

• Oh Kwang Il
• Kang Tae Wook
• Kim Sung Eun
• Park Mi Jeong
• Park Seong Mo
• Park Hyung-Il
• Lee Jae-Jin
• Lim In Gi

Classifications

• G06N3/08

  Learning methods · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• A61B1/0002

  provided with data storages · CPC title

• A61B1/00006

  of control signals · CPC title