Classification apparatus for detecting a state of a space with an integrated neural network, classification method, and computer readable medium storing a classification program for same

Patent metadata

Field	Value
Publication number	US-11475294-B2
Application number	US-201916386431-A
Country	US
Kind code	B2
Filing date	Apr 17, 2019
Priority date	Dec 6, 2016
Publication date	Oct 18, 2022
Grant date	Oct 18, 2022

Abstract

A classification apparatus includes: a specifying unit integrated with a neural network that has been trained to classify a state of a space using information indicating a light projection pattern and information indicating a light reception pattern; a light projection information acquisition unit configured to acquire information indicating a light projection pattern of light projected into a predetermined space, and output the acquired information to the specifying unit; and a light receiving unit configured to acquire information indicating a light reception pattern of light received from the predetermined space, and output the acquired information to the specifying unit, wherein the specifying unit outputs a classification result of classifying a state of the predetermined space, based on the information indicating the light projection pattern acquired by the light projection information acquisition unit and on the information indicating the light reception pattern of the light received by the light receiving unit.

First claim

The invention claimed is:

1. A classification apparatus comprising a communication interface, a sensor and a processor coupled to the communication interface and the sensor, the processor configured with a program to perform operations comprising:
detecting a state of a space indicating information about a detection target, with an integrated neural network trained to classify the state of the space using information received from the communication interface indicating a light projection pattern and information indicating a light reception pattern;
acquiring, from the sensor, the information indicating the light projection pattern of light projected into a predetermined space, and outputting the acquired information to the integrated neural network through the communication interface; and
acquiring, from the sensor through the communication interface, the information indicating the light reception pattern of light received from the predetermined space, and outputting the acquired information to the integrated neural network through the communication interface, wherein
the processor is configured with the program to perform operations such that detecting the state of the space comprises outputting, to the communication interface, a classification result of the classifying the state of the predetermined space, based on the acquired information indicating the light projection pattern and based on the acquired information indicating the light reception pattern of the light.

2. The classification apparatus according to claim 1 , wherein the processor is configured with the program to perform operations comprising operations such that:
acquiring the information indicating the light projection pattern comprises acquiring information indicating the light projection pattern of light projected onto the detection target within the predetermined space, and outputting the acquired information to the integrated neural network through the communication interface;
acquiring the information indicating a light reception pattern comprises acquiring a light reception pattern of the light projected onto the detection target within the predetermined space, and outputting the acquired light reception pattern to the integrated neural network through the communication interface; and
detecting of the space comprises outputting a classification result of classifying a presence or a state of the detection target within the predetermined space.

3. The classification apparatus according to claim 1 , wherein
the neural network is trained to classify the state of the space using information indicating the light projection pattern of light that is projected into the space that is given in advance, and information indicating the light reception pattern of light received from the space that is given in advance.

4. The classification apparatus according to claim 1 , wherein
the neural network is trained to classify the state of the space using information indicating a plurality of light projection patterns, and information indicating light reception patterns of the light that is received in response to projecting the plurality of light projection patterns.

5. The classification apparatus according to claim 1 , wherein
the neural network is trained to classify the state of the space using learning data in which the information indicating the light projection pattern, the information indicating the light reception pattern, and information indicating the state of the space are associated with each other.

6. The classification apparatus according to claim 1 , wherein
the neural network is trained to classify the state of the space using auxiliary information from an auxiliary sensor, in addition to the information indicating the light projection pattern and the information indicating the light reception pattern.

7. The classification apparatus according to claim 2 , wherein
the neural network is trained to classify the state of the space using information indicating the light projection pattern of light that is projected into the space that is given in advance, and information indicating the light reception pattern of light received from the space that is given in advance.

8. The classification apparatus according to claim 2 , wherein
the neural network is trained to classify the state of the space using information indicating a plurality of light projection patterns, and information indicating light reception patterns of light that is received in response to projecting the plurality of light projection patterns.

9. The classification apparatus according to claim 2 , wherein
the neural network is trained to classify the state of the space using learning data in which the information indicating the light projection pattern, the information indicating the light reception pattern, and information indicating the state of the space are associated with each other.

10. The classification apparatus according to claim 2 , wherein
the neural network is trained to classify the state of the space using auxiliary information from an auxiliary sensor, in addition to the information indicating the light projection pattern and the information indicating the light reception pattern.

11. The classification apparatus according to claim 3 , wherein
the neural network is trained to classify the state of the space using information indicating a plurality of light projection patterns, and information indicating light reception patterns of light that is received in response to projecting the plurality of light projection patterns.

12. The classification apparatus according to claim 3 , wherein
the neural network is trained to classify the state of the space using learning data in which the information indicating the light projection pattern, the information indicating the light reception pattern, and information indicating the state of the space are associated with each other.

13. The classification apparatus according to claim 3 , wherein
the neural network is trained to classify the state of the space using auxiliary information from an auxiliary sensor, in addition to the information indicating the light projection pattern and the information indicating the light reception pattern.

14. The classification apparatus according to claim 4 , wherein
the neural network is trained to classify the state of the space using learning data in which the information indicating the light projection pattern, the information indicating the light reception pattern, and information indicating the state of the space are associated with each other.

15. The classification apparatus according to claim 4 , wherein
the neural network is trained to classify the state of the space using auxiliary information from an auxiliary sensor, in addition to the information indicating the light projection pattern and the information indicating the light reception pattern.

16. The classification apparatus according to claim 5 , wherein
the neural network is trained to classify the state of the space using auxiliary information from an auxiliary sensor, in addition to the information indicating the light projection pattern and the information indicating the light reception pattern.

17. A classification method comprising:
detecting a state of a space with an integrated neural network trained to classify the state of the space using information received from a communication interface indicating a light projection pattern and information indicating a light reception pattern;
acquiring, from a sensor over the communication interface, information in

Assignees

• Omron Tateisi Electronics Co

Inventors

• Ando Tanichi

Classifications

• G06V10/141

  Control of illumination · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06V10/764 Primary

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

• G06N3/09 Primary

  Supervised learning · CPC title

• G06N3/088

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