JP2004275220A - Utterance meter, method of measuring quantity of utterance, program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an utterance meter capable of measuring the contents and situation of utterance of a user excluding the utterance the user does not intend, and a method of measuring the quantity of utterance. <P>SOLUTION: The utterance meter comprises a sound input means 1 for inputting sound, a voice recognition means 2 for recognizing the voice on the signal of the sound inputted by the sound input means 1, and an utterance quantity measuring means 3 comprising a classification database 6 with a plurality of words previously classified into a plurality of categories according to a prescribed standard, a classification means 4 for classifying the result of recognition from the voice recognition means by referring to the classification database 6 and a counter means 5 for counting the result of classification from the classification means 4 by category as the quantity of utterance. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a speech volume meter and a speech volume measurement method.
[0002]
[Prior art]
2. Description of the Related Art In recent years, as the social environment is changing to be called a stress society, interest in mental health is increasing. The human linguistic activity is used as an evaluation scale for mental health, dementia, autism, etc., and is used as a reference for mental therapy.
[0003]
A conventional device for evaluating linguistic activity converts a subject's vocal sound into an electric signal using a microphone, and integrates the time or amount of the electric signal transmitted, or both in the measurement time. Thus, it is possible to obtain, as objective data, the amount spoken by the subject within the measurement time (for example, see Patent Document 1).
[0004]
[Patent Document 1]
JP-A-5-56951
[0005]
[Problems to be solved by the invention]
However, in the related art, since the amount of speech of the user is simply counted based on the intensity level of the voice signal, the speech may be counted even when the user does not intend to speak, such as coughing, and speech activity may be counted. There was a problem in reliability for use such as measurement.
[0006]
In addition, it was not possible to obtain information necessary for mental treatment, such as the actual utterance content and the utterance amount for a specific word.
[0007]
In addition, it was not possible to measure the situation at the time of speech, such as whether he spoke to a person, whether it was a single word, or whether he was angry, and could not be used to evaluate psychological therapy for interpersonal relationships. .
[0008]
The present invention considers the problems of the above conventional utterance volume meter, removes utterances not intended by the user, and also measures the utterance content of the user and the situation at the time of utterance, and an utterance volume measurement method. It is intended to provide.
[0009]
[Means for Solving the Problems]
In order to achieve the above object, a first aspect of the present invention provides a sound input means (1) for inputting sound, a voice recognition means (2) for recognizing a sound signal input by the sound input means, An utterance amount meter including utterance amount measuring means (3) for measuring an utterance amount based on a result from the voice recognition means.
[0010]
In a second aspect of the present invention, the voice recognition means associates a plurality of voices not intended by the user with a plurality of voices predicted to be uttered by the user and a word represented by the voices, and records the voices in advance. The utterance meter according to the first aspect of the present invention, which has a registered voice database (202) and is a voice recognition unit that recognizes the sound signal from the recorded voice database.
[0011]
In a third aspect of the present invention, the voice recognition means includes a voice detection means (201) for detecting a voice waveform from the sound signal, and a voice waveform of the user and words represented by the voice waveform of the user. Associating and comparing a pre-recorded registered voice database (202) with a voice waveform registered in the registered voice database sequentially with a voice waveform input from the voice detecting means to calculate a similarity An arithmetic unit (203); and a recognition unit (204) for outputting a word corresponding to the speech waveform having the highest similarity as a recognition result. A classification database (6, 6 ′) classified into a plurality of categories according to criteria, a classification unit (4) for classifying words input from the recognition unit with reference to the classification database, A utterance meter according to a first aspect of the present invention, comprising counter means (5) for counting a result from the classification means as an utterance amount for each category.
[0012]
Further, the fourth invention further includes an emotion determination unit (407) for performing emotion determination based on the result from the voice detection unit or the recognition unit, and the classification unit includes The utterance meter according to the third aspect of the present invention, which is a classifying means for classifying the result from the voice recognition means in consideration of the result of (1).
[0013]
According to a fifth aspect of the present invention, the emotion determining means includes a power measuring means (401) for measuring a magnitude of a voice waveform input from the voice detecting means, or a voice input from the voice detecting means. A speed measuring unit (402) that measures and outputs an evaluation value of the utterance speed by performing frequency analysis of the waveform, or an emotion classifying unit that classifies and outputs an emotion represented by a word input from the recognition unit ( 404) and an emotion specifying means (403) for specifying the user's emotion using a result from the power measuring means, the speed measuring means, or the emotion classifying means. It is an utterance meter.
[0014]
The sixth aspect of the present invention further includes a physiological measurement unit (601) for measuring physiological information of the user, wherein the emotion determining unit determines an emotion when a change of a predetermined threshold or more occurs in the physiological information. And outputs an emotion determination result. If the physiological information does not change by a predetermined threshold value or more, the emotion determination unit outputs the previous emotion determination result without performing emotion determination. It is a meter.
[0015]
In a seventh aspect of the present invention, the physiological measuring means includes an infrared light emitting means (701) for irradiating the user's blood vessel with infrared light, the infrared light reflected by the blood vessel, and an intensity of the infrared light. An infrared light receiving means (702) for outputting a proportional signal, a pulse wave using the signal input from the infrared light receiving means, and a pulse wave detecting means (703) for outputting the pulse wave; and an input from the pulse wave detecting means A speech volume meter according to a sixth aspect of the present invention, which is a physiological measurement unit having a heart rate counting unit (704) for calculating and outputting a heart rate per unit time using the obtained pulse wave.
[0016]
An eighth aspect of the present invention further comprises a human body detecting means (801) for detecting whether or not there is a person in front of the user, wherein the utterance amount measuring means detects the human body detection information from the human body detecting means. The utterance amount meter according to the first aspect of the present invention, which is an utterance amount measurement unit that measures the added utterance amount.
[0017]
In a ninth aspect of the present invention, the human detecting means includes a pyroelectric sensor (901) for detecting radiant heat of the object, and a fluctuation detecting means for measuring a change in a temperature distribution detected by the pyroelectric sensor (901). 902) and a human body determining means (903) for determining that a target object is a human when there is a change amount equal to or more than a predetermined reference based on the result of the fluctuation detecting means.
[0018]
Further, a tenth aspect of the present invention provides a sound inputting step of inputting a sound, a voice recognizing step of recognizing a sound signal input in the sound inputting step, and an utterance based on a result from the voice recognizing step. And an utterance amount measuring step of measuring the amount.
[0019]
An eleventh aspect of the present invention is the speech recognition means according to the first, fourth, seventh or ninth aspect, wherein the speech recognition means recognizes a sound signal input by the sound input means. Based on the utterance amount measuring means for measuring the utterance amount, the voice detecting means, or the emotion determining means for performing emotion determination based on the result from the recognizing means, and the pulse using signals input from the infrared light receiving means. A pulse wave detecting means for forming and outputting a pulse, a heart rate per unit time is calculated using a pulse wave input from the pulse wave detecting means, and a heart rate counting means for outputting the pulse wave is detected by the pyroelectric sensor. A fluctuation detecting means for measuring a change amount of the temperature distribution, and a program for causing a computer to function as a human body judging means for judging an object as a human when there is a change amount equal to or more than a predetermined reference from the result of the fluctuation detecting means. It is.
[0020]
According to a twelfth aspect of the present invention, there is provided a recording medium carrying the program according to the eleventh aspect of the present invention, which is a recording medium usable by a computer.
[0021]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0022]
(Embodiment 1)
FIG. 1 is a block diagram of the utterance meter according to the first embodiment. The utterance meter according to the first embodiment includes a sound input unit 1 for inputting a user's voice. The apparatus further includes voice recognition means 2 for converting a voice signal into characters, and utterance amount measurement means 3 for counting the number of converted characters and measuring the amount of utterance.
[0023]
FIG. 2 shows a block diagram of the utterance amount measuring means 3. As shown in the figure, the utterance amount measuring means 3 has a classification database 6 that classifies characters in advance and stores them, and a classification unit 4 that classifies characters according to the classification stored in the classification database 6. are doing. Further, there is provided a counter means 5 for counting the number of characters recorded for each category as a user's utterance amount and notifying the user or the like.
[0024]
FIG. 3 shows a block diagram of the voice recognition means 2. As shown in the figure, the voice recognition unit 2 has a voice detection unit 201 that cuts out a voice section uttered by a user from a voice signal waveform and outputs a voice signal waveform of the section as a voice waveform. Here, a sound signal including noise input from the sound input means 1 is hereinafter referred to as an "audio signal waveform", which is obtained by removing noise from an audio signal waveform and cutting out an audio portion related to actual utterance. Is hereinafter referred to as a “voice waveform”.
[0025]
Further, a registered voice database 202 in which a voice waveform corresponding to a word predicted to be spoken by the user is input, and the input voice waveform and a voice waveform registered in the registered voice database 202 are compared. A comparison operation unit 203 for detecting a similarity is provided for each registered voice waveform. Further, there is provided a recognition unit 204 that detects a speech waveform having the highest similarity, determines a corresponding word from the registered speech database 202, and performs character conversion.
[0026]
An utterance amount measuring method using the utterance meter according to the first embodiment having the above-described configuration will be described below.
[0027]
First, the sound input unit 1 converts a voice input by a user into a voice signal waveform and outputs the voice signal waveform to the voice recognition unit 2. Subsequently, the voice detection unit 201 cuts out a waveform portion where the voice signal waveform satisfies a certain standard as a voice section actually uttered by the user, and outputs the voice signal waveform of the section to the comparison calculation unit 203 as a voice waveform. Here, the fixed reference for cutting out a voice section includes, for example, that the power of a signal waveform in a frequency band of 1 kHz or less, which is generally a human voice band, is equal to or higher than a certain level. Further, the above-described criterion may be applied after a physical vocalization operation is detected by a vibration sensor or a myoelectric sensor attached to a user's throat or chin.
[0028]
In the registered voice database 202, a voice waveform corresponding to a word predicted to be uttered by the user and the word are registered in advance in association with each other. Then, the comparison operation unit 203 sequentially compares the audio waveform input from the audio detection unit 201 with the audio waveform registered in the registered audio database 202, and performs similarity for each audio waveform registered in the registered audio database 202. The degree is calculated and output to the recognition means 204.
[0029]
Here, as a comparison method, the two voice waveforms to be compared may be the total sum of the power components at each frequency after frequency analysis such as Fourier transform, or the cepstrum feature amount and the mel code that are further subjected to the polar coordinate conversion after the frequency analysis. A method of performing DP matching in consideration of temporal expansion and contraction in the cepstrum feature amount may be used. Further, a sound waveform registered in the registered sound database 202 may be used as a comparison element used in the comparison operation means 203, for example, a power component of each frequency for the purpose of efficient comparison operation.
[0030]
In addition, in the voice waveform registered in the registered voice database 202, a voice waveform when the user does not intend to utter, such as a user's coughing and humming, is registered, and the corresponding word is registered as “unintended utterance”. deep. This makes it possible to distinguish unintended utterances.
[0031]
Subsequently, the recognizing unit 204 detects a speech waveform having the highest similarity among the similarities of the respective speech waveforms input from the comparison operation unit 203, and determines a corresponding word from the registered speech database 202. Then, character conversion of the audio waveform is performed. If there is no large difference between the similarities, the input voice is determined to be noise and no character conversion is performed. Alternatively, the character may be converted to “noise”.
[0032]
Next, the classification unit 4 classifies the words input from the recognition unit 204 during the use period according to a predetermined category classification as shown in FIG. Here, the category classification is, for example, words in which the voices have the same content such as "Good morning" or "Good morning."
[0033]
Finally, the counter unit 5 counts the number of words recorded for each category by the classifying unit 4 during the use period as the amount of utterance of the user, and indicates a numerical value for each category to the user, a doctor, or the like.
[0034]
This makes it possible to distinguish utterances not intended by the user, measure the amount of utterance for each utterance content, and evaluate the original amount of language activity necessary for psychotherapy.
[0035]
In the first embodiment, the voice recognition unit of the present invention converts a voice waveform into a word by using a registered voice waveform and words “Ohayo”, “Ohayosao” or the like indicated by the voice. , "Flower", "blooming", etc., or one character such as "a", "e" may be subjected to voice recognition, and the resulting words may be classified.
[0036]
In addition, noise sound other than voice is also input to the sound input means of the present invention, and the voice of the present invention also includes unintended occurrences such as cough and characters.
[0037]
(Embodiment 2)
FIG. 5 is a block diagram of the speech meter according to the second embodiment. The utterance meter according to the second embodiment has the same basic configuration as that of the first embodiment, but differs in that an emotion determination unit 407 for adding an emotion at the time of utterance and performing classification is also provided. Therefore, the description will focus on this difference. The same components as those in the first embodiment are denoted by the same reference numerals.
[0038]
The emotion determination unit 407 of the utterance meter according to the second embodiment includes a power measurement unit 401 that measures the frequency power of a speech waveform. Further, there is provided a speed measuring means 402 for evaluating the speed of the voice waveform from the frequency of the voice waveform. Further, the apparatus includes an emotion classification database 405 in which words and emotions represented by the words are previously associated and stored, and an emotion classification unit 404 for performing emotion classification according to the emotion classification database 405.
[0039]
Further, it has an emotion identification database 406 in which the results of the frequency power, speed, and emotion classification and the emotion determination result using them are associated with each other, and an emotion identification unit 403 that performs emotion identification according to the emotion identification database 406.
[0040]
Further, the utterance meter according to the first embodiment has the classification database 6 in which words and their category classifications are associated, but the utterance meter according to the second embodiment emits the words. It has a classification database 6 'to which the emotion of time is further added.
[0041]
An utterance amount measuring method using the utterance amount meter according to the second embodiment having the above configuration will be described below.
[0042]
First, as in the first embodiment, the voice detection unit 201 cuts out a voice section actually uttered by the user from the voice signal waveform input from the sound input unit 1, and converts the voice signal waveform of the section into voice. The waveform is output to the power measurement unit 401 as a waveform. Next, the power measuring unit 401 measures the frequency power of the input audio waveform, determines the input volume level, and outputs it to the emotion specifying unit 403.
[0043]
In the same manner as described above, the audio waveform is cut out by the audio detection means 201, and the output audio waveform is subjected to frequency analysis such as Fourier transform. The speech speed is evaluated on the assumption that the speech speed is high when the distribution has a large high-frequency component as a whole, and the speech speed is low when the low-frequency component is large. At this time, the standard utterance speed of the user is measured in advance as an evaluation criterion, and the evaluation value of the utterance speed is determined by comparing the frequency component distribution with the reference value 0, and is output to the emotion specifying means 403. Here, when it is determined that there are many high-frequency components and the utterance speed is fast, a positive value may be used, or a negative value may be used.
[0044]
The emotion classification unit 404 classifies the words converted by the voice recognition unit 2 into words representing anger, sadness, joy, or the like according to the emotion classification database 405 registered in advance. It is output to the specifying means 403. Here, FIG. 6 shows an example of the emotion classification database 405. As shown in the figure, words and emotions corresponding to the words are registered.
[0045]
Next, based on the input sound volume level output from the power measurement unit 401, the evaluation value of the utterance speed output from the speed measurement unit 402, and the emotion classification output from the emotion classification unit 404, according to the emotion identification database 406, The emotion identification means 403 performs emotion identification such as normality, joy, anger, sadness and the like at the time of utterance, and outputs it to the classification means 4. In the emotion identification database 406, for example, a determination table shown in FIG. 7 is recorded. In this determination table, when it is determined that "anger", the voice recognition means 2 recognizes the word as expressing anger, measures the input volume level as being large or normal, and determines that the generation speed is fast or normal. Only when it is done, the judgment of “anger” is made.
[0046]
Subsequently, the classification unit 4 adds the emotion classification result to each category held in the classification database 6 ′ to perform classification. Here, FIG. 8 shows an example of the classification database 6 ′ according to the second embodiment. As shown in the figure, it is possible to measure the utterance amount obtained by further dividing one category for each emotion.
[0047]
Lastly, the counter means 5 counts the number of words recorded for each category and for each emotion classification by the classification means 4 during the use period as a user's utterance amount, and provides the user and doctor etc. for each category and Shows numerical values for each emotion classification.
[0048]
This makes it possible to measure the amount of speech for each emotional state when the user speaks, and to provide more detailed information that can be used for psychotherapy to doctors and caregivers.
[0049]
In the second embodiment, the emotion determination unit of the present invention includes a power measurement unit 401, a speed measurement unit 402, an emotion classification unit 404, an emotion identification unit 403, and an emotion identification database 406. The power measuring unit 401, the speed measuring unit 402, and the emotion classifying unit 404 are arbitrarily selected, and the emotion determination is performed using only those results. In other words, it only needs to be an emotion determination unit that determines the emotion of the sound signal.
[0050]
In the second embodiment, the emotion classification unit of the present invention includes the emotion classification unit 404 and the emotion classification database 405. However, the emotion classification database 405 may be incorporated in the emotion classification unit 404.
[0051]
In the second embodiment, the emotion specifying means of the present invention includes the emotion specifying means 403 and the emotion specifying database 406. However, the emotion specifying database 406 may be incorporated in the emotion specifying means 403.
[0052]
Further, the emotion classification means of the present invention may also serve as the emotion identification means.
[0053]
(Embodiment 3)
FIG. 9 is a block diagram of a speech meter according to the third embodiment. The utterance meter according to the third embodiment has the same basic configuration as that of the second embodiment, except that the utterance meter measures the physiological information of the user and makes an emotion determination in consideration of a change in the physiological information. . Therefore, this difference will be mainly described. The same components as those in the second embodiment are denoted by the same reference numerals.
[0054]
The utterance meter according to the third embodiment further includes a physiological measurement unit 601 for measuring physiological information of the user. Here, a configuration diagram of the physiological measurement unit 601 is shown in FIG. As shown in the figure, the physiological measurement means 601 has an infrared light emitting means 701 for irradiating an infrared ray to a blood vessel and an infrared light receiving means 702 for reflecting a blood vessel and outputting a signal corresponding to the intensity of the incident infrared light. ing. Here, as shown in FIG. 10, the infrared light emitting means 701 and the infrared light receiving means 702 irradiate a user's earlobe or a blood vessel in a finger with a clip structure or a ring-shaped structure, and reflected light thereof is transmitted to the infrared light receiving means 702. It is structured to be incident. Also, a pulse wave is formed using a signal input from the infrared light receiving unit 702, and a pulse wave detecting unit 703 that outputs the pulse wave, and a heart rate per unit time is calculated using the pulse wave input from the pulse wave detecting unit 703. Is calculated and output to the emotion specifying means 403.
[0055]
An utterance amount measuring method using the utterance meter configured as described above will be described below.
[0056]
First, as shown in FIG. 10, an infrared light emitting means 701 and an infrared light receiving means 702 are attached to an earlobe or a finger of a user by a clip structure or a ring-shaped structure, and infrared rays emitted from the infrared light emitting means 701 are transmitted to the earlobe or the finger. And the reflected light is incident on the infrared light receiving means 702. The infrared light receiving means 702 outputs a signal proportional to the intensity of the incident infrared light to the pulse wave detecting means 703.
[0057]
Here, the intensity of the infrared light incident on the infrared light receiving means 702 is inversely proportional to the absorbance of the blood vessel, but the blood vessel contracts and expands as the heart beats, and the absorbance changes accordingly. The pulse wave detecting means 703 reconstructs the pulse wave of the heart from the above-mentioned change in the absorbance of the blood vessel, and outputs the reconstructed pulse wave to the heart rate counting means 704.
[0058]
Subsequently, the heart rate counting unit 704 calculates a period of one cycle of the pulse wave from the pulse wave data, calculates a heart rate per unit time, and outputs the calculated heart rate to the emotion specifying unit 403.
[0059]
Next, the emotion specifying unit 403 evaluates an input from the physiological measurement unit 601 based on the assumption that “when a change in emotion occurs, a change in physiological information also occurs.” In this case, based on the results of the power measurement unit 401, the speed measurement unit 402, and the emotion classification unit 404, emotion determination such as normal, joy, anger, sadness, and the like at the time of speech is performed according to the determination table illustrated in FIG. Is output to the classification means 4.
[0060]
If the amount of change in the physiological information is equal to or less than a certain reference, the previous emotion determination result is output to the classification unit 4. The classification unit 4 adds and records and retains the result of emotion classification for each of the held categories, and the counter unit 5 stores the words recorded for each category and each emotion classification by the classification unit 4 during the use period. Is counted as the user's utterance amount, and the numerical value for each category and each emotion classification is shown to the user, doctor, and the like.
[0061]
This makes it possible to more accurately evaluate the utterance of the user as the amount of utterance for each emotional state at the time of utterance.
[0062]
Although the heart rate is raised in the third embodiment as the physiological information of the present invention, the amount of perspiration, the body temperature, the amount of blinking of the eyes, and the like may be detected.
[0063]
(Embodiment 4)
FIG. 11 is a block diagram of the utterance meter according to the fourth embodiment. The utterance meter according to the fourth embodiment has the same basic configuration as that of the first embodiment, but includes a human body detection unit 801 that detects whether or not a person is speaking when speaking. Is different. Therefore, this difference will be mainly described. The same components as those in the first embodiment are denoted by the same reference numerals.
[0064]
FIG. 12 shows a configuration diagram of the human body detection means 801. As shown in the figure, the human body detecting means 801 has a pyroelectric sensor 901 for detecting radiant heat from an object. Further, there are provided a fluctuation detecting means 902 for measuring a change amount of the temperature distribution detected by the pyroelectric sensor 901 and a human body determining means 903 for judging whether or not the target object is a human based on the result of the change amount.
[0065]
Further, the utterance meter according to the first embodiment has the classification database 6 in which words and their categorization are associated, but the utterance meter according to the fourth embodiment emits the words. It has a classification database 6 "to which the presence or absence of human detection as to whether or not the user is speaking to a person is sometimes added.
[0066]
An utterance amount measuring method using the utterance meter configured as described above will be described below.
[0067]
The pyroelectric sensor 901 is attached so as to face the front of the user as shown in FIG. 12, detects infrared rays due to radiant heat from the target object by a pyroelectric effect, and outputs its intensity.
[0068]
Next, the fluctuation detecting means 902 obtains the area of the temperature distribution from the normal human body temperature of 34 degrees to 40 degrees from the input intensity from the pyroelectric sensor 901 and calculates the amount of change of the area per unit time. It outputs to the human body determination means 903 as the amount of change in the temperature distribution.
[0069]
Next, the human body determination unit 903 has a reference width of the amount of change in the temperature distribution when a person is talking face-to-face in a normal state, and the input from the fluctuation detection unit 902 falls within the reference width. If so, the object in front is determined to be a human body and output to the classifying means 4.
[0070]
Next, the classification means 4 adds the result of the human body detection means 801 to each category held in the classification database 6 "to perform classification. Here, FIG. 13 shows the classification database 6" in the fourth embodiment. Here is an example. As shown in the figure, the database is a database in which the presence or absence of human detection is added to each word in one category.
[0071]
Finally, the counter means 5 counts the number of words recorded by the classification means 4 for each category and each time when there is or is not a person in front during the use period as a user's utterance amount, And doctors, etc., are shown for each category and the numerical value of whether or not it is interpersonal utterance.
[0072]
As a result, it is possible to measure the amount of utterance by distinguishing the utterance of the user into the case of utterance to a person and the case of not, and obtain information on mental activities related to interpersonal relationships. It can provide doctors and caregivers with useful information for treatment of social phobia and social stress.
[0073]
In the fourth embodiment, a pyroelectric sensor is used as the human body detecting means of the present invention. However, for example, an optical sensor such as a CCD sensor may be used, in other words, whether a person is in front of the user. It only needs to be a person detecting means for determining whether or not it is.
[0074]
Note that the program of the present invention is a program for causing a computer to execute the functions of all or a part of the utterance meter of the present invention (or an apparatus, an element, or the like), and cooperates with the computer. This is a working program.
[0075]
Further, the recording medium of the present invention carries a program for causing a computer to execute all or a part of functions of all or a part of the utterance meter of the present invention (or an apparatus, an element, or the like). It is a recording medium readable by a computer, and the read program is a recording medium that executes the function in cooperation with the computer.
[0076]
One usage form of the program of the present invention may be a form in which the program is recorded on a computer-readable recording medium and operates in cooperation with the computer.
[0077]
One use form of the program of the present invention may be a form in which the program is transmitted through a transmission medium, read by a computer, and operates in cooperation with the computer.
[0078]
The recording medium includes a ROM and the like, and the transmission medium includes a transmission medium such as the Internet, light, radio waves, and sound waves.
[0079]
Further, the computer of the present invention described above is not limited to pure hardware such as a CPU, but may include firmware, an OS, and peripheral devices.
[0080]
Note that, as described above, the configuration of the present invention may be realized by software or hardware.
[0081]
【The invention's effect】
As is clear from the above description, it is possible to provide an utterance meter capable of distinguishing utterances not intended by the user.
[0082]
Further, it is possible to provide an utterance meter capable of measuring the amount of utterance for each emotional state, and measuring the amount of utterance separately when there is a person.
[Brief description of the drawings]
FIG. 1 is a block diagram of an utterance meter according to Embodiment 1 of the present invention.
FIG. 2 is a block diagram of an utterance meter according to the first embodiment of the present invention.
FIG. 3 is a configuration diagram of a voice recognition unit according to the first embodiment of the present invention.
FIG. 4 is a diagram showing an example of category classification of a classification database according to the first embodiment of the present invention;
FIG. 5 is a block diagram of a speech volume meter according to Embodiment 2 of the present invention.
FIG. 6 is a diagram showing an example of emotion classification in an emotion classification database according to Embodiment 2 of the present invention.
FIG. 7 is a diagram showing an example of an emotion identification table in an emotion identification database according to Embodiment 2 of the present invention.
FIG. 8 is a diagram showing an example of category classification of a classification database according to the second embodiment of the present invention;
FIG. 9 is a block diagram of an utterance meter according to Embodiment 3 of the present invention.
FIG. 10 is a configuration diagram of a physiological measurement unit according to Embodiment 3 of the present invention.
FIG. 11 is a block diagram of an utterance meter according to Embodiment 4 of the present invention.
FIG. 12 is a configuration diagram of a human body detection unit according to a fourth embodiment of the present invention.
FIG. 13 is a diagram showing an example of category classification in a classification database according to a fourth embodiment of the present invention.
[Explanation of symbols]
One sound input means
2 voice recognition means
3 utterance volume measurement means
4 classification means
5 counter means
6 classification database

Claims (12)

Sound input means for inputting sound,
Voice recognition means for recognizing the voice of the sound signal input by the sound input means,
An utterance amount meter comprising: an utterance amount measurement unit that measures an utterance amount based on a result from the voice recognition unit. The voice recognition means,
A plurality of voices not intended by the user and a plurality of voices predicted to be uttered by the user, and associating the words represented by the voices, having a pre-recorded registered voice database,
2. The utterance meter according to claim 1, wherein the utterance meter is a voice recognition unit for recognizing the sound signal from the recorded voice database. The voice recognition means,
Sound detection means for detecting a sound waveform from the sound signal,
Correlating the words represented by the user's voice waveform and the user's voice waveform, a pre-recorded registered voice database,
Comparison arithmetic means for sequentially comparing the audio waveform registered in the registered audio database with the audio waveform input from the audio detection means, and calculating a similarity;
A recognition unit that outputs a word corresponding to the speech waveform having the highest similarity as a recognition result,
The utterance amount measuring means,
A classification database in which a plurality of words are classified in advance into a plurality of categories according to predetermined criteria,
Classification means for classifying the words input from the recognition means with reference to the classification database,
2. The utterance meter according to claim 1, further comprising counter means for counting a result from the classification means as an utterance amount for each category. The voice detection means, or further comprising an emotion determination means for performing emotion determination based on the result from the recognition means,
The utterance meter according to claim 3, wherein the classification unit is a classification unit that classifies a result from the voice recognition unit in consideration of a result from the emotion determination unit. The emotion determination means,
Power measuring means for measuring the magnitude of the sound waveform input from the sound detecting means,
Or, by performing a frequency analysis of the voice waveform input from the voice detection means to measure the evaluation value of the speech speed, speed measurement means to output,
Or, emotion classification means to classify the emotion represented by the words input from the recognition means, and output,
The utterance meter according to claim 4, further comprising: an emotion identification unit that identifies an emotion of the user by using a result from the power measurement unit, the speed measurement unit, or the emotion classification unit. Further comprising a physiological measurement means for measuring physiological information of the user,
The emotion determination means performs an emotion determination when a change equal to or more than a predetermined threshold occurs in the physiological information, outputs an emotion determination result, and performs an emotion determination when a change equal to or more than the predetermined threshold does not occur in the physiological information. 5. The utterance meter according to claim 4, wherein the utterance meter is an emotion determination unit that outputs a previous emotion determination result. The physiological measurement means,
Infrared light emitting means for irradiating the user's blood vessels with infrared light,
Infrared light receiving means for receiving the infrared light reflected by the blood vessel and outputting a signal proportional to the intensity of the infrared light,
A pulse wave is configured using a signal input from the infrared light receiving unit, and a pulse wave detecting unit that outputs the pulse wave,
7. The vocal meter according to claim 6, wherein the stimulus meter is a physiological measurement unit having a heart rate counting unit that calculates and outputs a heart rate per unit time using the pulse wave input from the pulse wave detecting unit. Further comprising a human body detection means for detecting whether there is a person in front of the user,
The utterance amount meter according to claim 1, wherein the utterance amount measurement unit is an utterance amount measurement unit that measures an utterance amount in consideration of human body detection information from the human body detection unit. The human detection means, a pyroelectric sensor for detecting radiant heat of the object,
Fluctuation detecting means for measuring the amount of change in the temperature distribution detected by the pyroelectric sensor,
9. The utterance meter according to claim 8, further comprising: a human body determination unit that determines that the target object is a human when a change amount equal to or more than a predetermined reference is found from the result of the fluctuation detection unit. A sound input process of inputting sound,
A voice recognition step of recognizing a voice of a sound signal input in the sound input step,
An utterance amount measuring step of measuring an utterance amount based on a result from the voice recognition step. Speech recognition means for recognizing a sound signal input by said sound input means of the utterance meter according to claim 1, 4, 7, or 9.
An utterance amount measuring unit that measures an utterance amount based on the voice recognition unit,
The voice detection unit, or an emotion determination unit that performs an emotion determination based on the result from the recognition unit,
A pulse wave is configured using a signal input from the infrared light receiving unit, and a pulse wave detecting unit that outputs the pulse wave.
Using a pulse wave input from the pulse wave detection means, calculates a heart rate per unit time, heart rate counting means to output,
Fluctuation detecting means for measuring the amount of change in the temperature distribution detected by the pyroelectric sensor,
And a program for causing a computer to function as a human body determination unit that determines that an object is a human when there is a change amount equal to or more than a predetermined reference based on a result of the fluctuation detection unit. A recording medium carrying the program according to claim 11, wherein the recording medium is usable by a computer.

Images