Identification apparatus, identification method, identification processing program, generation apparatus, generation method, and generation processing program

Patent metadata

Field	Value
Publication number	US-12078613-B2
Application number	US-202017602483-A
Country	US
Kind code	B2
Filing date	Apr 9, 2020
Priority date	Apr 12, 2019
Publication date	Sep 3, 2024
Grant date	Sep 3, 2024

Abstract

An identification device according to one embodiment includes a vibration generation unit that generates, by a vibration generator, first vibrations to be provided to a three-dimensional object to be identified having an integrated structure; an acquisition unit that acquires, from a vibration detector, a detection signal corresponding to a second vibration that has propagated inside the three-dimensional object among the first vibrations provided to the three-dimensional object; a feature quantity generation unit that generates a feature quantity indicating a frequency characteristic of the second vibration, based on the acquired detection signal; and an identification unit that identifies the three-dimensional object to be identified, based on a feature quantity stored in a storage device in which the feature quantity indicating a frequency characteristic based on a vibration that has propagated inside a previously identified three-dimensional object is stored and on the feature quantity generated by the feature quantity generation unit.

First claim

The invention claimed is:

1. An identification device comprising:
a memory configured to store a program; and
a processor configured to execute the program so as to:
cause a vibration generator to generate first vibrations to be provided to a three-dimensional object to be identified having an integrated structure;
acquire, from a vibration detector, a detection signal corresponding to a second vibration that has propagated inside the three-dimensional object among the first vibrations provided to the three-dimensional object;
generate a feature quantity indicating a frequency characteristic of the second vibration, based on the acquired detection signal; and
identify the three-dimensional object to be identified based on feature quantities stored in the memory and the generated feature quantity, each of the stored feature quantities indicating a frequency characteristic based on a vibration that has propagated inside each of previously identified three-dimensional objects,

wherein an identifier uniquely assigned to the vibration generator that has generated the first vibrations relating to identification and an identifier uniquely assigned to the vibration detector that has detected the second vibration relating to identification are stored in the memory in association with the stored feature quantities of the previously identified three-dimensional objects, and
the processor is further configured to:
extract corresponding feature quantities from the stored feature quantities that are respectively associated with the identifier assigned to the vibration generator and the identifier assigned to the vibration detector; and
identify the three-dimensional object to be identified based on the extracted corresponding feature quantities and the generated feature quantity.

2. The identification device according to claim 1 ,
wherein the vibration generator is configured with a plurality of vibration generators, and the vibration detector is configured with a plurality of vibration detectors,
information indicating a contact surface of each of the plurality of vibration generators that contacts the three-dimensional object to be identified, information indicating a contact surface of each of the plurality of vibration detectors that contacts the three-dimensional object to be identified, and information indicating a surface of the three-dimensional object to be identified are stored in the memory, and
the processor is further configured to:
based on the information stored in the memory, select a first vibration generator to be used for generating the first vibrations from the plurality of vibration generators, the first vibration generator having the contact surface adapted to the surface of the three-dimensional object to be identified; and
based on the information stored in the memory, select a first vibration detector to be used for acquiring the detection signal from the plurality of vibration detectors, the first vibration detector having the contact surface adapted to the surface of the three-dimensional object to be identified.

3. The identification device according to claim 1 ,
wherein a model, into which the generated feature quantity for the three-dimensional object to be identified is input and which outputs a value based on a difference between the stored feature quantity of at least one of the previously identified three-dimensional objects and the generated feature quantity of the three-dimensional object to be identified in association with an identifier uniquely assigned to the at least one of the previously identified three-dimensional objects, is stored in the memory,
the model is learned based on the generated feature quantity, and
wherein the processor is further configured to:
input the generated feature quantity for the three-dimensional object to be identified into the model; and
identify the three-dimensional object to be identified by determining that the identifier output in association with a value indicating a highest degree of similarity to the generated feature quantity of the three-dimensional object to be identified among values output from the model is an identifier of the three-dimensional object to be identified.

4. The identification device according to claim 3 ,
wherein the identifier assigned to the vibration generator and the identifier assigned to the vibration detector are further input into the model, and
the value based on the difference between the stored feature quantity of at least one of the previously identified three-dimensional objects and the generated feature quantity of the three-dimensional object to be identified is output from the model in association with the identifiers assigned to the vibration generator and the vibration detector.

5. The identification device according to claim 3 ,
wherein the model is learned based on supervisory data in which the generated feature quantity for the three-dimensional object to be identified and an identifier uniquely assigned to the three-dimensional object to be identified are associated with each other.

6. An identification method for causing a processor to execute a process, the identification method comprising executing on the processor the steps of:
causing a vibration generator to generate first vibrations to be provided to a three-dimensional object to be identified having an integrated structure;
acquiring, from a vibration detector, a detection signal corresponding to a second vibration that has propagated inside the three-dimensional object among the first vibrations provided to the three-dimensional object;
generating a feature quantity indicating a frequency characteristic of the second vibration, based on the acquired detection signal; and
identifying the three-dimensional object to be identified based on feature quantities stored in a memory and the generated feature quantity, each of the stored feature quantities indicating a frequency characteristic based on a vibration that has propagated inside each of previously identified three-dimensional objects,
wherein an identifier uniquely assigned to the vibration generator that has generated the first vibrations relating to identification and an identifier uniquely assigned to the vibration detector that has detected the second vibration relating to identification are stored in the memory in association with the stored feature quantities of the previously identified three-dimensional objects, and
the processor is further configured to:
extract corresponding feature quantities from the stored feature quantities that are respectively associated with the identifier assigned to the vibration generator and the identifier assigned to the vibration detector; and
identify the three-dimensional object to be identified based on the extracted corresponding feature quantities and the generated feature quantity.

7. The identification method according to claim 6 ,
wherein a model, into which the generated feature quantity for the three-dimensional object to be identified is input and which outputs a value based on a difference between the stored feature quantity of at least one of the previously identified three-dimensional objects and the generated feature quantity of the three-dimensional object to be identified in association with an identifier uniquely assigned to the at least one of the previously identified three-dimensional objects, is stored in the memory,
the model is learned based on the generated feature quantity, and
the processor is further configured:
input the generated feature quantity generated for the three-dimensional object to be identified into the model; and
identify the three-dimensional object to be identified by determining that the identifier output in association with a value

Assignees

• Nippon Telegraph & Telephone

Inventors

• Kubo Yuki

Classifications

• G01N29/4418 Primary

  with a model, e.g. best-fit, regression analysis · CPC title

• G01N29/46

  by spectral analysis, e.g. Fourier analysis {or wavelet analysis (spectral signal processing per se G06F17/14)} · CPC title

• G01N2291/023

  Solids · CPC title

• G01N2291/024

  Mixtures · CPC title

• G01N29/2437

  Piezoelectric probes · CPC title