# US 2025/0095644 A1 · Estimation method, recording medium, and estimation device

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0095644 A1
- Document ID: US20250095644A1
- Status: Published application
- Kind: A1
- Application number: 18/292,214
- Filed: 2022-07-19
- Published: 2025-03-20
- Earliest claimed: 2021-08-06
- Assignees: Panasonic Intellectual Property Management Co., Ltd.
- Applicants: Panasonic Intellectual Property Management Co., Ltd.
- Inventors: Masashi ISHIMARU; Yoshihiro MATSUMURA; Hiroki KAMEI; Jakusei KIYOSAKI; Yaming ZHANG; Yuki KUNUGIDA; Takuya YAMAMURA
- Web page: https://patentgrep.com/patent/US20250095644A1

## CPC

- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L 25/66: for extracting parameters related to health condition (detecting or measuring for diagnostic purposes A61B5/00)
- A61B 5/16: Devices for psychotechnics (using teaching or educational appliances G09B1/00 - G09B7/00); Testing reaction times {; Devices for evaluating the psychological state}
- A61B 5/4803: {Speech analysis specially adapted for diagnostic purposes}
- A61B 5/7267: {involving training the classification device}
- A61B 10/00: Instruments for taking body samples for diagnostic purposes (for taking samples of blood A61B5/15); Other methods or instruments for diagnosis, e.g. for vaccination diagnosis, sex determination or ovulation-period determination; Throat striking implements
- G10L 25/51: for comparison or discrimination
- G10L 21/003: Changing voice quality, e.g. pitch or formants
- G10L 2015/225: {Feedback of the input speech}
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)

## Abstract

An estimation method includes: obtaining a first voice feature group of a plurality of persons who speak a first language; obtaining a second voice feature group of a plurality of persons who speak a second language; obtaining a voice feature of a subject; correcting the voice feature of the subject according to a relationship between the first voice feature group and the second voice feature group; estimating, from the voice feature of the subject that has been corrected, an oral function or a cognitive function of the subject by using an estimation process for an oral function or a cognitive function based on the second language; and outputting a result of estimation of the oral function or the cognitive function of the subject.

## Claims

### Claim 1

An estimation method executed by an estimation device that estimates an oral function or a cognitive function, the estimation method comprising:
obtaining a voice feature group of a plurality of persons who speak a first language, by analyzing voices obtained by prompting the plurality of persons who speak the first language to speak a first language sentence that is in the first language and is similar in pronunciation to a second language sentence spoken in a second language different from the first language;
obtaining a voice feature group of a plurality of persons who speak the second language, by analyzing voices obtained by prompting the plurality of persons who speak the second language to speak the second language sentence;
obtaining a voice feature of a subject who speaks the first language, by analyzing a voice obtained by prompting the subject to speak the first language sentence;
correcting the voice feature of the subject according to a relationship between the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language;
estimating, from the voice feature of the subject that has been corrected, an oral function or a cognitive function of the subject by using an estimation process for an oral function or a cognitive function based on the second language; and
outputting a result of estimation of the oral function or the cognitive function of the subject.

### Claim 2 (depends on 1)

The estimation method according to claim 1, further comprising:
generating a correction expression for matching a distribution of the voice feature group of the plurality of persons who speak the first language to a distribution of the voice feature group of the plurality of persons who speak the second language, according to the relationship between the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language,
wherein in the correcting, the voice feature of the subject is corrected using the correction expression.

### Claim 3 (depends on 2)

The estimation method according to claim 2,
wherein a voice feature of a voice obtained from the first language sentence being spoken includes a plurality of types of voice features,
a voice feature of a voice obtained from the second language sentence being spoken includes a plurality of types of voice features,
in the generating of the correction expression, the correction expression is generated for each of the plurality of types of voice features, and
in the correcting, the correction expression is selected according to a type of a voice feature, and the voice feature of the subject is corrected using the correction expression selected.

### Claim 4 (depends on 2)

The estimation method according to claim 2,
wherein the correction expression includes, for each of percentile sections of a frequency distribution of the voice feature group of the plurality of persons who speak the first language and a frequency distribution of the voice feature group of the plurality of persons who speak the second language, a coefficient for correcting a frequency of the frequency distribution of the voice feature group of the plurality of persons who speak the first language, and
in the correcting, the voice feature of the subject is corrected using the coefficient corresponding to a percentile section that includes the voice feature of the subject.

### Claim 5 (depends on 2)

The estimation method according to claim 2,
wherein the correction expression includes means and standard deviations of frequency distributions of the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language, and
in the correcting, the voice feature of the subject is corrected using the means and the standard deviations.

### Claim 6 (depends on 5)

The estimation method according to claim 5,
wherein the generating of the correction expression includes:
performing logarithmic transformation on the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language to obtain a logarithmically-transformed voice feature group of the plurality of persons who speak the first language and a logarithmically-transformed voice feature group of the plurality of persons who speak the second language; and
generating the correction expression for matching a distribution of the logarithmically-transformed voice feature group of the plurality of persons who speak the first language to a distribution of the logarithmically-transformed voice feature group of the plurality of persons who speak the second language, according to a relationship between the logarithmically-transformed voice feature group of the plurality of persons who speak the first language and the logarithmically-transformed voice feature group of the plurality of persons who speak the second language,
the correction expression includes means and standard deviations of frequency distributions of the logarithmically-transformed voice feature group of the plurality of persons who speak the first language and the logarithmically-transformed voice feature group of the plurality of persons who speak the second language, and
the correcting includes:
performing logarithmic transformation on the voice feature of the subject to obtain a logarithmically-transformed voice feature of the subject;
correcting the logarithmically-transformed voice feature of the subject by using the means and the standard deviations; and
performing exponential transformation on the logarithmically-transformed voice feature of the subject that has been corrected.

### Claim 7 (depends on 1)

The estimation method according to claim 1, further comprising:
generating a trained model for approximating a voice feature of a voice obtained from the first language sentence being spoken to a voice feature of a voice obtained from the second language sentence being spoken, by teaching the relationship between the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language,
wherein in the correcting, the voice feature of the subject is corrected using the trained model.

### Claim 8 (depends on 7)

The estimation method according to claim 7,
wherein the voice feature of the voice obtained from the first language sentence being spoken includes a plurality of types of voice features,
the voice feature of the voice obtained from the second language sentence being spoken includes a plurality of types of voice features,
in the generating of the trained model, the trained model is generated for each of the plurality of types of voice features, and
in the correcting, the trained model is selected according to a type of a voice feature, and the voice feature of the subject is corrected using the trained model selected.

### Claim 9 (depends on 1)

The estimation method according to claim 1, further comprising:
providing a suggestion regarding the oral function or the cognitive function for the subject by checking the result of estimation against predetermined data.

### Claim 10 (depends on 1)

A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to execute the estimation method according to claim 1.

### Claim 11

An estimation device that estimates an oral function or a cognitive function, the estimation device comprising:
a first voice feature group obtainer that obtains a voice feature group of a plurality of persons who speak a first language, by analyzing voices obtained by prompting the plurality of persons who speak the first language to speak a first language sentence that is in the first language and is similar in pronunciation to a second language sentence spoken in a second language different from the first language;
a second voice feature group obtainer that obtains a voice feature group of a plurality of persons who speak the second language, by analyzing voices obtained by prompting the plurality of persons who speak the second language to speak the second language sentence;
a subject voice feature obtainer that obtains a voice feature of a subject who speaks the first language, by analyzing a voice obtained by prompting the subject to speak the first language sentence;
a corrector that corrects the voice feature of the subject according to a relationship between the voice feature group of the plurality of persons who speak the first language and the voice feature group of the plurality of persons who speak the second language;
an estimator that estimates, from the voice feature of the subject that has been corrected, an oral function or a cognitive function of the subject by using an estimation process for an oral function or a cognitive function based on the second language; and
an outputter that outputs a result of estimation of the oral function or the cognitive function of the subject.
