JP2004243023A - Laughter detector, information processor and laughter detection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a laughter detector and an information processor capable of accurately detecting the number of times of laughters over a long period of time by detecting the number of times of the laughters by discriminating the laughters and speeches from sound, eliminating the need of a place for installing an imaging devices and miniaturizing the laughter detector. <P>SOLUTION: The laughter detector is provided with a sound acquisition part for acquiring the sound, a signal extraction part for extracting only the signals of a fixed frequency band from the sound acquired by the sound acquisition part, a judgement part for judging whether or not the sound acquired by the sound acquisition part is the laughter on the basis of the signals extracted by the signal extraction part, and a judged result output part for outputting a judged result in the judgement part. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a laughter and utterance section detection device that distinguishes laughter from an utterance section and detects the number of laughs.
[0002]
Another aspect of the present invention relates to an information processing apparatus that acquires health information regarding mental health based on laughter information including the number of laughs.
[0003]
[Prior art]
As a first conventional technique, there is a method of recognizing laughter based on a dynamic change in a hue of a cheek of a complexion under the laughing emotion in communication using the skin temperature of the face as a physiological index (see Non-Patent Document 1). ).
[0004]
As a second related art, there is a blood pressure management system that stores a blood pressure value measured by a sphygmomanometer and manages the blood pressure value. On the basis of such accumulated blood pressure value data, a doctor diagnoses the health condition of the subject, or the subject himself pays attention to the health (see Non-Patent Document 2).
[0005]
[Non-patent document 1]
Journal of Human Interface Society, Vol3, No. "Dynamic Analysis of Facial Color in Laughter by Simultaneous Measurement of Facial Image and Facial Skin Temperature" described on pages 2, 23-30 (2001)
[Non-patent document 2]
Matsushita Electric Works, Ltd. Website Internet
<URL: http: // dmedia. mew. co. jp / wellness / kestuatsu / b-ew284. html>
[0006]
[Problems to be solved by the invention]
However, the first conventional technique determines laughter based on a dynamic model that averages variations in the hue of a face between individuals.According to the variation in the hue of a face, it is possible to accurately determine laughter. Did not. In addition, it is necessary to fix the photographing device for photographing the face and the user, and there is a problem that it is difficult to detect laughter for a long time by restricting the user.
[0007]
In the second conventional technique, a doctor or the subject himself / herself diagnoses the current health condition based on the accumulated blood pressure value, and the blood pressure value is normal, high blood pressure, low blood pressure based on the blood pressure value. It outputs a simple comment such as blood pressure, and does not serve as an index for determining the state of mental health.
[0008]
[Means for Solving the Problems]
In order to solve the above-described problems, the present invention is a laughter detection device that detects the number of laughter based on sound. Since a place for installation is not required and the laughter detection device can be downsized, the number of times of laughter can be accurately detected over a long period of time.
[0009]
Another aspect of the present invention is an information processing apparatus that configures mental health information based on laughter information including the number of laughs. And health information of the heart can be diagnosed. The health information is information on mental health, and may include a state of stress, a mental state, and the like, or a state of a physical function resulting from other mental stress such as a brain, heart, and internal organs.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Note that, in this embodiment, components using the same reference numerals, steps in a flowchart, and the like have the same function, and therefore, description thereof may be omitted once.
[0011]
(Embodiment 1)
[0012]
FIG. 1 is a block diagram illustrating a configuration of the laughter detection device according to the present embodiment. The laughter detection device includes a start reception unit 101, a sound acquisition unit 102, a determination start reception unit 103, a signal extraction unit 104, a determination unit 105, and a determination result output unit 106.
[0013]
The start accepting unit 101 accepts an input instruction to instruct the start of voice acquisition. The start receiving unit 101 can be generally realized by means such as a switch, a button, a touch panel, and the like, which are input by a person. Further, the start receiving unit 101 may be realized by, for example, a voice recognition unit or the like that recognizes a voice of a start uttered by a person.
[0014]
The sound acquisition unit 102 acquires a sound. The sound is information about a voice, for example, a voice, a vocal cord vibration generated by vibrating a vocal cord, and the like. Acquisition of such a sound can usually be realized by a bone conduction microphone, a voice microphone, or the like. The term “acquisition” is a concept that includes receiving from an external device different from the laughter detection device.
[0015]
Further, the sound acquisition unit 102 records the acquired sound signal on a recording medium built in the sound acquisition unit 102. The sound acquisition unit 102 is generally realized by hardware (dedicated circuit) including a recording medium for recording a sound signal, but may be realized by software. This recording medium may be a nonvolatile recording medium or a volatile recording medium.
[0016]
The determination start receiving unit 103 receives an input instruction. The input instruction is an instruction to start a process of extracting a signal in a frequency band of a vowel sound from a sound signal recorded by the sound acquisition unit 102. The determination start accepting unit 103 is generally realized by means such as a switch, a button, a touch panel, and the like, which are input by a person. Further, the determination start receiving unit 103 includes, for example, a voice recognition unit or the like for recognizing a voice “start of signal extraction” uttered by a human, and a timer built therein, and is fixed from the time when the start receiving unit 101 receives an input instruction. It can also be realized by means for automatically starting a process of extracting a signal of a frequency band of a vowel sound after a lapse of time (for example, after a lapse of 12 hours).
[0017]
The signal extraction unit 104 reads the sound signal recorded by the sound acquisition unit 102. The signal extraction unit 104 extracts a signal in a certain sound frequency band from the sound signal. This signal extraction can be realized by, for example, performing a wavelet transform process on a sound signal and extracting a frequency band of a vowel sound. Details will be described later. The signal extracting unit 104 is generally realized by hardware (dedicated circuit) including a recording medium for recording a sound signal, but may be realized by software. This recording medium may be a nonvolatile recording medium or a volatile recording medium.
[0018]
The determination unit 105 distinguishes between a laughing section and an uttering section from the signal extracted by the signal extracting unit 104, determines only the laughing section, and detects the number of laughs. Further, the number of laughs per unit time from the time when the start receiving unit 101 receives the input instruction to the time when the determination start receiving unit 103 receives the input instruction is calculated. Here, the utterance means a normal conversation. Details will be described later. The determination unit 105 is usually configured by software, but may be realized by hardware (dedicated circuit).
[0019]
The determination result report output unit 106 outputs the number of laughs detected by the determination unit 105 and the number of laughs per unit time. The number of laughs may be the number of laughs per day or the number of laughs per month. The output is preferably displayed on a display or the like, but may be output from a speaker or printed on a printer, and is a concept including transmission. The determination result output unit 106 can be realized by, for example, a display and its driver software. Further, the determination result output unit 106 may not include a display. In such a case, the determination result output unit 106 instructs the external display to output the number of laughs and the number of laughs per unit time. Furthermore, when outputting the number of times of laughter by sound, the determination result output unit 106 may include a speaker and its driver software. When the determination result output unit 106 transmits information or the like regarding the number of times of laughter to a device different from the laughter detection device, the determination result output unit 106 includes, for example, a network card and its driver software. realizable.
[0020]
Hereinafter, the operation of the laughter detection device according to the present embodiment will be described with reference to the flowcharts of FIGS. 2 and 3. The flowchart of FIG. 2 describes an operation of acquiring a sound, extracting a signal in a frequency band of a vowel sound, calculating an envelope from the extracted signal, and detecting an amplitude peak equal to or more than a certain value. . The flow chart of FIG. 3 distinguishes a laughter section and an utterance section from the detected amplitude peak, detects the number of laughs and the number of laughs per unit time, and outputs the number of laughs and the number of laughs per unit time. The following describes the operation to be performed.
[0021]
(Step S201) The start receiving unit 101 determines whether or not an input instruction for starting sound acquisition has been received. If an input instruction has been received, the process proceeds to step S202, and if not, the process returns to step S201.
[0022]
(Step S202) The sound acquisition unit 102 determines whether a sound has been acquired. If a sound has been obtained, the process proceeds to step S203. If no sound has been obtained, the process returns to step S202.
[0023]
(Step S203) The sound acquisition unit 102 records the sound signal acquired in Step 202 and the acquired time in a memory.
[0024]
(Step S204) The determination start receiving unit 103 determines whether or not an input instruction to start a process of extracting a signal in a frequency band of a vowel sound from the sound signal recorded in Step 203 has been received. If an input instruction has been received, the process proceeds to step S205, and if not, the process returns to step S202.
[0025]
(Step S205) The signal extraction unit 104 determines whether or not the sound signal recorded in the memory in Step S203 exists. If there is a sound signal, the process proceeds to step S206. If there is no sound signal, the procedure goes to step S209. After reading, the signal extraction unit 104 increments the pointer of the memory by one to read the next sound signal and the acquisition time.
[0026]
(Step S206) The signal extraction unit 104 reads the sound signal and the acquisition time from the memory.
[0027]
(Step S207) The signal extraction unit 104 performs wavelet transform on the sound signal read in step 206. Wavelet transform means that a sound signal is divided into two parts, a low frequency band and a high frequency band, by a two-band filter bank, and the low frequency band is divided into a low frequency band and a high frequency band. That is, it is repeatedly divided into two. A band-divided filter bank is a divided filter derived from a wavelet function.
[0028]
(Step S208) The signal extraction unit 104 extracts a signal in the frequency band of the vowel sound, and records the signal and the acquisition time corresponding to the signal in the memory. Here, the frequency band of the vowel sound is a frequency band from 0 to 1000 Hz. A signal in the frequency band of the vowel sound recorded in the memory is called a vowel frequency band signal. After recording the vowel frequency band signal and the acquisition time corresponding to the signal in the memory, the signal extraction unit 104 returns to step 205.
[0029]
(Step S209) The determining unit 105 determines whether or not the vowel frequency band signal recorded in the memory in Step S208 exists. If there is a vowel frequency band signal, the procedure goes to step S210. If there is no vowel frequency band signal, the procedure goes to step S214. After reading, the determination unit 105 counts up the memory pointer by one to read the next sound signal and the acquisition time.
[0030]
(Step S210) The determination unit 105 reads the vowel frequency band signal and the acquisition time from the memory.
[0031]
(Step S211) The determination unit 105 determines the absolute value of the amplitude of the vowel frequency band signal read from the memory.
[0032]
(Step S212) The determining unit 105 performs an LPF (Low Pass Filer) process on the absolute value of the amplitude obtained in step 211 to obtain an envelope of the amplitude.
[0033]
(Step S213) The determination unit 105 records the amplitude envelope data obtained in step 212 and the corresponding acquisition time in the memory. The recorded data of the amplitude envelope is used as amplitude envelope data.
[0034]
(Step S214) The determination unit 105 initializes the amplitude peak acquisition time T (m).
[0035]
(Step S215) The determination unit 105 sets the counter i and the counter m to 1 each. The i-th of the counter i means the recording order in which the amplitude envelope data was recorded at the i-th. That is, the amplitude envelope data when the counter i is 1 means the amplitude envelope data recorded first. The m-th counter m indicates a recording order in which the acquisition time corresponding to the amplitude peak is recorded at the m-th time at the amplitude peak acquisition time T (m) when the amplitude peak is equal to or greater than a certain value. That is, the amplitude peak acquisition time T (m) when the counter m is 1 means the acquisition time corresponding to the amplitude peak when the amplitude peak recorded first is equal to or more than a certain value.
[0036]
(Step S216) The determining unit 105 determines whether or not the i-th amplitude envelope data recorded in the memory in step S213 exists. If the i-th amplitude envelope data exists, the process proceeds to step S217. If the i-th amplitude envelope data does not exist, the procedure goes to step S301.
[0037]
(Step S217) The determination unit 105 reads the amplitude envelope data and the acquisition time from the memory.
[0038]
(Step S218) The determining unit 105 detects a peak of the amplitude from the amplitude envelope data read in step 217.
[0039]
(Step S219) The determination unit 105 determines whether the peak of the amplitude obtained in step 217 is about 0.05 or more. If the amplitude peak is about 0.05 or more, go to step 220. If the peak of the amplitude is not greater than about 0.05, the process proceeds to step 222. The reason for determining whether the peak of the amplitude is about 0.05 or more is to remove miscellaneous signals other than laughter.
[0040]
(Step S220) The determination unit 105 records the acquisition time corresponding to the amplitude peak of about 0.05 or more determined in step 219 at the amplitude peak acquisition time T (m).
[0041]
(Step S221) The determining unit 105 increments the counter m by one.
[0042]
(Step S222) The determining unit 105 increments the counter i by one. The determination unit 105 returns to step 216 after counting up.
[0043]
(Step S301) The determination unit 105 initializes the number of laughs N and the number of laughs per unit time.
[0044]
(Step S302) The determination unit 105 sets the counter m to 1. The counter m is used to read the m-th recorded amplitude peak acquisition time T (m).
[0045]
(Step S303) The determination unit 105 determines whether or not the (m + 3) th amplitude peak acquisition time T (m + 3) recorded in step S213 exists. If the (m + 3) th amplitude peak acquisition time T (m + 3) exists, the process proceeds to step S304. If the (i + 3) th amplitude peak acquisition time T (m + 3) does not exist, the process proceeds to step S315.
[0046]
(Step S304) The determination unit 105 reads the amplitude peak acquisition time T (m), the amplitude peak acquisition time T (m + 1), the amplitude peak acquisition time T (m + 2), and the amplitude peak acquisition time T (m + 3).
[0047]
(Step S305) The determination unit 105 obtains a difference between the amplitude peak acquisition time T (m) and the amplitude peak acquisition time T (m + 1), and sets the difference as X1. Next, a difference between the amplitude peak acquisition time T (m + 1) and the amplitude peak acquisition time T (m + 2) is obtained, and the difference is set to X2. Next, a difference between the amplitude peak acquisition time T (m + 2) and the amplitude peak acquisition time T (m + 3) is obtained, and the difference is set to X3.
[0048]
(Step S306) The determination unit 105 determines the maximum value of X1, X2, and X3, and sets the maximum value as Y1. Next, the minimum value of X1, X2, and X3 is obtained, and the minimum value is set as Y2.
[0049]
(Step S307) The determination unit 105 determines whether the difference between Y1 and Y2 is about 30 ms or less. If the difference between Y1 and Y2 is about 30 ms or less, the process proceeds to step 308. If the difference between Y1 and Y2 is not less than about 30 ms, go to step 310.
[0050]
(Step S308) The determination unit 105 adds 1 to the number N of laughter. Here, if the difference between Y1 and Y2 is about 30 ms or less, it is determined that laughter is occurring if the time interval between the peaks of three consecutive amplitudes is about 30 m or less from the experimental results. This is because it can be determined that laughter has occurred.
[0051]
(Step S309) The determination unit 105 increments the counter m by four. Next, the determination unit 105 returns to Step 303.
[0052]
(Step S310) The determining unit 105 determines whether Y1 is the same as X1. If Y1 is the same as X1, go to step 311. If Y1 is not the same as X1, go to step 312.
[0053]
(Step S311) The determining unit 105 increments the counter m by one. Next, the determination unit 105 returns to Step 303.
[0054]
(Step S312) The determination unit 105 determines whether Y1 is the same as X2. If Y1 is the same as X2, go to step 313. If Y1 is not the same as X1, go to step 314.
[0055]
(Step S313) The determination unit 105 increments the counter m by two. Next, the determination unit 105 returns to Step 303.
[0056]
(Step S314) The determination unit 105 increments the counter m by three. Next, the determination unit 105 returns to Step 303.
[0057]
(Step S315) The determination unit 105 calculates the number of laughs per unit time based on the number of laughs N.
[0058]
(Step S316) The determination result output unit 106 outputs the number of laughs N and the number of laughs per unit time. After this output, the operation of the laughter detection device ends.
[0059]
The operation in which the laugh detection device detects the number of laughs and the number of laughs per unit time and outputs the number of laughs and the number of laughs per unit time has been described above with reference to the flowcharts of FIGS. 2 and 3.
[0060]
Hereinafter, a specific operation of the laughter detection device according to the present embodiment will be described. The sound acquisition unit of the laughter detection device has a bone conduction microphone 401. FIG. 4 shows a bone conduction microphone 401 and a bone conduction microphone 402 when worn by a user. The bone conduction microphone may be mounted so as to be in close contact with the larynx, or may be mounted so as to be in close contact with the external auditory meatus or the top of the head. Further, the bone conduction microphone 401 may transmit the acquired sound signal to the sound acquisition unit via a wire or may transmit the acquired sound signal by wireless communication means.
[0061]
Now, the user is wearing the bone conduction microphone 502 closely on the larynx and watching a television. FIG. 5 shows a laughter detection device 501, a bone conduction microphone 502, a push button 503 of a start reception unit, a push button 504 of a judgment start reception unit, and a liquid crystal display 505 of a judgment result output unit.
[0062]
Push button 503 of the start receiving unit receives an input instruction for instructing start of sound acquisition.
[0063]
Next, the bone conduction microphone 502 acquires the vocal cord vibration and converts it into an audio signal to acquire a sound. When the bone transmission microphone acquires the voice, the voice of the user can be directly obtained, and the influence of the surrounding noise can be reduced. Further, the acquired audio signal is associated with the time at which the audio signal was acquired, and is recorded in a memory built in the sound acquisition unit.
[0064]
Next, the push button 504 of the determination start receiving unit receives an input instruction to start a process of extracting a signal of a frequency band of a vowel sound based on the acquired sound signal.
[0065]
Next, the signal extraction unit reads the audio signal and the acquisition time recorded in the memory. Next, a signal in the frequency band of the vowel voice is extracted from the voice signal. In the extraction of the signal, first, an audio signal in a frequency band of 0 to 8000 Hz is extracted. Next, the extracted audio signal of 0 to 8000 Hz is subjected to wavelet transform processing to extract an audio signal of a frequency band of 0 to 1000 Hz which is a frequency band of the vowel sound. The frequency band of the vowel sound is defined as a vowel frequency band signal. The wavelet transform is to divide an audio signal into two parts, a low frequency band and a high frequency band, using a two-band filter bank, and further divide the low frequency band into a low frequency band and a high frequency band. That is, it is repeatedly divided into two. A band-divided filter bank is a divided filter derived from a wavelet function. That is, the audio signal of 0 to 8000 Hz is divided into two of a frequency band of 0 to 4000 Hz and a frequency band of 4000 to 8000 Hz by a two-band filter bank, and the audio signal of the frequency band of 0 to 4000 Hz is The frequency band from 0 to 2000 Hz and the frequency band from 2000 to 4000 Hz are divided into two, and the audio signal from the frequency band from 0 to 2000 Hz is further divided into the frequency band from 0 to 1000 Hz and the frequency band from 1000 to 2000 Hz. Finally, an audio signal in a frequency band of 0 to 1000 Hz is extracted.
[0066]
FIG. 6 shows the result of the wavelet transform. According to FIG. 6, an audio signal 601 in a frequency band from 0 to 8000 Hz, an audio signal 602 in a frequency band from 0 to 1000 Hz, an audio signal 603 in a frequency band from 1000 to 2000 Hz, and an audio signal 604 and 4000 in a frequency band from 2000 to 4000 Hz. 5 shows an audio signal 605 in a frequency band from 8000 Hz to 8000 Hz.
[0067]
In the present embodiment, the audio signal in the frequency band of 0 to 1000 Hz is extracted from the audio signal by using the means of the wavelet transform, but the audio signal is subjected to the low-frequency band pass filter processing of 0 to 1000 Hz. Thus, the audio signal in the frequency band from 0 to 1000 Hz may be extracted, or may be extracted using other means.
[0068]
Next, the signal extraction unit sequentially records the extracted vowel frequency band signal and the corresponding acquisition time in a memory built in the signal extraction unit.
[0069]
Next, the determination unit sequentially reads the vowel frequency band signal and the corresponding acquisition time from a memory built in the signal extraction unit.
[0070]
Next, the determination unit calculates the absolute value of the amplitude of the read vowel frequency band signal, performs LPF processing of 5 to 10 Hz, and obtains amplitude envelope data of the vowel frequency band signal. FIG. 7 shows a time series change of the amplitude envelope data of the vowel frequency band signal. FIG. 7 shows an amplitude envelope 701 of a laughing voice waveform and an amplitude envelope 702 of an uttering voice waveform.
[0071]
In the present embodiment, the absolute value of the signal is calculated, and the amplitude envelope data is obtained by performing LPF processing at 5 to 10 Hz. You may ask.
[0072]
Next, the determining unit records the obtained amplitude envelope data and the corresponding acquisition time in the memory.
[0073]
Next, the determining unit detects a peak of the amplitude from the read amplitude envelope data. If the peak of the amplitude is about 0.05 or more, the acquisition time corresponding to the amplitude peak of about 0.05 or more is recorded at the amplitude peak acquisition time T (m). Then, the counter m is incremented by one. The reason why the peak of the amplitude is about 0.05 or more is to remove miscellaneous signals other than laughter. In addition, about 0.05 may be another value.
[0074]
Next, the determination unit counts up the counter i by one, reads the existing i-th amplitude envelope data, detects the amplitude peak, and if the amplitude peak is about 0.05 or more, the amplitude peak Recording the acquisition time corresponding to the amplitude peak of about 0.05 or more at the acquisition time T (m) is repeated. This process is repeated until the amplitude envelope data no longer exists.
[0075]
Next, the judgment unit detects the number of laughs N. The determination unit reads the amplitude peak acquisition time T (m + 1), the amplitude peak acquisition time T (m + 2), and the amplitude peak acquisition time T (m + 3) from the amplitude peak acquisition time T (m).
[0076]
Next, the determining unit calculates the period of the continuous amplitude peak. The difference between the amplitude peak acquisition time T (m) and the amplitude peak acquisition time T (m + 1) is obtained, and the difference is defined as X1. The difference between the amplitude peak acquisition time T (m + 1) and the amplitude peak acquisition time T (m + 2) is obtained, and the difference is defined as X2. The difference between the amplitude peak acquisition time T (m + 2) and the amplitude peak acquisition time T (m + 3) is obtained, and the difference is defined as X3. Next, the maximum value of X1, X2, and X3 is obtained, and the maximum value is set to Y1. The minimum value of X1, X2, X3 is obtained, and the minimum value is set to Y2. According to FIG. 7, X1 indicates the first interval 7011, X2 indicates the second interval 7012, X3 indicates the third interval 7013, and the first interval 7021 and the second interval 7022 in the case of speech.
[0077]
Next, the determining unit determines whether or not the cycle of the peak of the continuous amplitude is within a certain range. Preferably, the period of the laughter amplitude peak is typically in the range of about 80 to about 120 ms. The cycle of the amplitude peak of laughter in the range of about 80 to about 120 ms means the range of the cycle when the magnitude of laughter is adjusted to some extent. The cycle of the amplitude peak of laughter fluctuates due to individual differences. When laughing large, the cycle is in a large range, for example, in the range of 150 to 190 ms, and when laughing small, the cycle is small. . Further, when the amplitude peak of laughter is repeated three times in succession, the range of the variation of the interval between the amplitude peaks is preferably about 30 ms. This about 30 ms was obtained from the measurement result of the repetition period of laughter and conversation. The value of about 30 ms may be another value. FIG. 8 shows experimental results obtained by measuring the period of laughter and the period of conversation. FIG. 9 shows variations in the cycle of laughter and conversation. This variation is derived from the experimental results shown in FIG. Also, from the result of this variation, it can be seen that the repetition periods of laughter and conversation are different, and a certain range 901 for distinguishing laughter from conversation is obtained. That is, the difference between Y1 and Y2 is obtained, and if the difference is within about 30 ms, it is determined that laughter has been performed, and 1 is added to the number N of laughter. Then, the counter m is incremented by 4 and the determination of whether or not the next continuous amplitude peak is laughing is repeated.
[0078]
When the difference between Y1 and Y2 is not within about 30 ms, if Y1 is X1, the determination unit increments the counter m by one. If Y1 is X2, the counter m is incremented by two. If Y1 is X3, the counter m is incremented by three. In this process, if the variation of the periodic interval between the amplitude peaks of three consecutive laughs does not fall within about 30 ms, the cycle of the three consecutive laughs is repeated from the next periodic interval of the maximum periodic interval. This is to determine the variation in the interval time.
[0079]
Next, the determination unit reads the amplitude peak acquisition time until the m + 3th amplitude peak acquisition time T (m + 3) recorded in the memory no longer exists, and repeats the operation of determining whether or not laughing has been performed. The number of laughs N is detected.
[0080]
Next, the determining unit divides the number of laughs N by the time from the time when the push button 503 receives the input instruction to the time when the push button 504 of the determination start receiving unit receives the input instruction, and Calculate the number of laughs.
[0081]
Next, the liquid crystal display 505 of the judgment result report output unit outputs the number N of laughs detected by the judgment unit and the number of laughs per unit time. FIG. 10 shows a display example of the number of laughs N output on the liquid crystal display and the number of laughs per unit time. According to FIG. 10, the number of laughs is 30 times, the number of laughs per unit time is 2, the detection date is February 7, the detection start time AM 8:00, and the detection end time PM 11:00 are displayed simultaneously. The detection date can be realized by obtaining the date on which the number of laughs is detected by the laughter detection device having the calendar function means. Further, the detection start time and the detection end time can be realized by displaying the time received by the start reception unit and the time received by the determination start reception unit.
[0082]
As described above, according to the present embodiment, a laughter detection device that can determine a laughter interval and an utterance interval from an audio signal, detect the number of laughter times, and output the number of laughter times, Thereby, the number of laughs can be detected accurately and for a long time without errors in the detection of the number of laughs due to individual differences. In addition, by providing the computer with the "laughter" detection function of the present invention, "laughter" is accurately recognized, and the computer can realize communication close to that of a human. In addition, by accurately detecting the number of laughs, it is useful for evaluation and production of games, entertainment, and the like. In addition, since the utterance section and the laughter section can be accurately determined, a profession that needs to manage the utterance time, for example, a singer or an announcer, can easily manage the utterance time. In addition, if the number of laughs is small, doctors and users are advised that stress may accumulate, develop mental illness such as depression, decrease physical function caused by mental factors, develop illness, etc. And the like can be easily determined based on the number of laughs. In addition, it has been reported as an experimental result that NK cells having an immune function are activated by laughing, and an accurate number of laughs is obtained using a laugh detection device to measure the NK cell activity rate. It is useful.
[0083]
Note that, in the present embodiment, as means for distinguishing between a laughter and an utterance section, the determination unit obtains an envelope of the vowel frequency band signal, and determines a period of the amplitude peak when the amplitude peak of the envelope is equal to or more than a certain value. Has been determined to be within a certain range, but means for determining from the shape of the amplitude waveform of the envelope may be used. That is, according to FIG. 7, the amplitude waveform of the laughter amplitude envelope has a substantially sinusoidal waveform. The determination unit records the signal having the substantially sinusoidal waveform in a memory in advance, compares the signal of the amplitude waveform of the amplitude envelope of the acquired voice with the signal having the substantially sinusoidal waveform, and determines the amplitude of the voice. If a waveform that approximates the shape of a sine waveform can be detected from the envelope waveform, it is determined that the section is a laughing interval. The means for detecting the shape of the substantially sinusoidal waveform includes, for example, a signal value at which the amplitude of the laughter amplitude waveform and the amplitude waveform of the audio amplitude envelope peak, and a signal at two points near the amplitude peak in time. Find the value. Next, the signal values at the three locations are compared, and when the signal values at the three locations match or are approximately similar, a substantially sinusoidal waveform shape can be detected. The means for detecting the shape of the substantially sinusoidal waveform may be another means.
[0084]
Alternatively, means for judging from the fundamental frequency (voice pitch) of the laughing voice may be used. The determination unit stores, for example, a standard fundamental frequency that is a fundamental frequency of a standard laughing voice, compares the standard fundamental frequency with the frequency of the acquired audio signal, and determines the frequency of the acquired audio signal. Is determined to be a laugh when it matches or almost approximates the standard fundamental frequency.
[0085]
In the case where the determination is made by combining two or more of the determination means according to the present embodiment, the means for determining from the shape of the amplitude waveform of the envelope, and the means for determining from the frequency (voice pitch) of the laughter voice, etc. The laughter section can be detected with higher accuracy.
[0086]
Further, in the present embodiment, the result determination output unit is a liquid crystal display, which displays the number of laughs and the number of laughs per unit time by numbers, but the number of laughs by voice and the number of laughs per unit time May be notified or printed. Further, a mode in which the number of laughs is transmitted to another device may be used. Such output control is the same in other embodiments.
[0087]
Furthermore, the operation of the laughter detection device described in the present embodiment may be realized by software, and the software may be placed on a server, for example, and the software may be distributed by software download. Further, the software may be recorded on a recording medium such as a CD-ROM and distributed. This is the same in all other embodiments.
[0088]
A program when the operation in the present embodiment is realized by software is as follows. In the computer, a sound acquisition step of acquiring sound, a signal extraction step of extracting only a signal in a certain frequency band from the sound acquired in the sound acquisition step, and a sound acquisition step based on the signal extracted in the signal extraction step. This is a program for executing a determination step of determining whether the acquired sound is laughter and a determination result output step of outputting a determination result in the determination step.
[0089]
(Embodiment 2)
FIG. 11 is a block diagram illustrating a configuration of an information processing system including an information processing device, a laughter detection device, and a health information display device according to the present embodiment. The laughter detection device 1110 includes a start reception unit 101, a sound acquisition unit 102, a determination start reception unit 103, a signal extraction unit 104, a determination unit 105, and a determination result output unit 1111. Further, the information processing device 1120 includes a laughter information receiving unit 1121, a health information forming unit 1122, and a health information output unit 1123. In addition, the health information display device 1130 includes a health information receiving unit 1131 and an information output unit 1132. The health information display device 1130 is a predetermined device identified by a user identifier.
[0090]
FIG. 12 is a block diagram illustrating a configuration of the health information configuration unit. The health information configuration unit 1210 has a health information acquisition unit 1211 and a health information storage unit 1212.
[0091]
Laugh detection device 1110 has the same configuration as that of the first embodiment, and detects the number of laughs and the number of laughs per unit time by the same operation. Then, laughter information including the laugh information and the user identifier is transmitted. This transmission is performed by the determination result output unit 1111. The determination result output unit 1111 includes, for example, a network card and its driver software, and can be realized by transmitting laughter information to the information processing device. The laugh information refers to information about laughter, and is, for example, the number of laughs or the number of laughs per unit time. The user identifier is information for identifying a user. The user identifier is, for example, an identification number, a name, or the like. Further, attribute information of the user according to the user identifier, for example, other attribute information such as age, gender, and past illnesses may be stored in a recording medium incorporated in the information processing device, and the laughter detection device May be stored in a built-in recording medium. The data structure of the user identifier does not matter. The information of the user identifier is recorded on a recording medium included in the laughter detection device. Further, the information of the user identifier may be recorded on a (external) recording medium different from the laughter detection device. This recording medium may be a nonvolatile recording medium or a volatile recording medium.
[0092]
The laughter information receiving unit 1121 of the information processing device 1120 receives laughter information. The laughter information receiving unit 1121 includes, for example, a network card and its driver software, and can be realized by receiving laughter information.
[0093]
The health information acquisition unit 1212 acquires health information from the health information stored in the health information storage unit 1211 based on the laughter information received by the laughter information reception unit 1121. The health information acquisition unit 1212 is usually configured by software, but may be realized by hardware (dedicated circuit).
[0094]
The health information storage unit 1211 stores health information related to mental health. The non-volatile recording medium is suitable for the health information storage unit 1211. However, the health information storage unit 1211 may be a volatile recording medium. Specifically, the health information storage unit 1211 is preferably a DVD (Digital Versatile Disc) or the like.
[0095]
The health information output unit 1123 outputs the health information acquired by the health information configuration unit 1122. The health information output unit 1123 is composed of, for example, a network card and its driver software, and can be realized by transmitting health information. Further, display on a display or the like is preferable, but sound output from a speaker or printing on a printer may be used. The health information output unit 1123 can be realized by, for example, a display and its driver software. Also, the health information output unit 1123 may not include a display. In such a case, the health information output unit 1123 instructs to output health information to an external display. Furthermore, when outputting the health information by sound, the health information output unit 1123 may include a speaker and its driver software.
[0096]
Health information receiving section 1131 receives health information. The health information receiving unit 1131 is composed of, for example, a network card and its driver software, and can be realized by receiving health information.
[0097]
The information output unit 1132 outputs the health information received by the health information receiving unit 1132. This output is preferably displayed on a display or the like, but may be audio output from a speaker or printing on a printer. The information output unit 1132 can be realized by, for example, a display and its driver software. The information output unit 1132 may not include a display. In such a case, the information output unit 1132 instructs the external display to output health information. Further, when outputting the health information by sound, the information output unit 1132 may include a speaker and its driver software.
[0098]
Hereinafter, a specific operation of the information processing apparatus according to the present embodiment will be described.
[0099]
FIG. 13 shows an information processing device 1301, a laughter detection device 1302, and a health information display device 1303. According to FIG. 13, the information processing device 1301, the laughter detection device 1302, and the health information display device 1303 are all connected via the Internet.
[0100]
The laughter detection device 1302 detects the number of laughs and the number of laughs per unit time by the same operation as in the first embodiment. Next, the determination result output unit of the laughter detection device 1302 transmits laughter information including the number of laughs, the number of laughs per unit time, a user identifier, and the like to the information processing device. FIG. 14 shows an example of laughter information data.
According to FIG. 14, the data of the laughter information includes an identification number “1055”, a name “Hanako Keihan”, a detection date, the number of laughs per unit time, the number of laughs, and the like. The data structure of the laughter information does not matter.
[0101]
In the present embodiment, the determination result output unit of the laughter detection device transmits one laughter information. However, the laughter frequency is detected a predetermined number of times (for example, the laughter frequency is detected 10 times). After that, or after a predetermined period (for example, one week) has elapsed, the laughter information may be transmitted. In addition, the laughter information of a plurality of users may be transmitted. In this case, the determination result output unit transmits the user identifier, the number of laughs corresponding to the user identifier, and the number of laughs per unit time as laughter information.
[0102]
The laughter information receiving unit receives the laughter information, and the health information obtaining unit obtains an identification number from the laughter information, and sequentially stores the laughter information in a memory allocated according to the identification number.
[0103]
Next, the health information acquiring means sequentially reads the number of laughs per unit time (for example, from January 1 to January 31) in a single month from the data of the laughter information stored in the memory. Calculate the average value of the number of laughs per unit time in a single month.
[0104]
Next, the health information acquiring means searches for data including the age, gender, name, etc. stored in a memory built in the information processing device in advance using the identification number as a key, and retrieves the age, gender, To get. Next, the health information acquisition means searches the health information data stored in the health information storage means using the average number of laughs per unit time, age and gender as keys, and retrieves the corresponding health information data. get. FIG. 15 shows that, using the identification number “1055” as a key, the data 1501 including the age, gender, name, and the like is searched to acquire the age and gender that matches the identification number “1055”. FIG. 16 shows a search of the health information data 1602 stored in the health information storage means 1603 using “average number of laughs, age, and gender per unit time” as a key, and searching for age, gender, and average unit time. This indicates that data of health information that matches the number of laughs is acquired.
[0105]
According to FIG. 15, using the identification number “1055” as a key, the data 1501 including the age, gender, name, and the like is searched, and the corresponding age “65” and gender “female” are acquired. According to FIG. 16, the health information data 1602 is searched using the average number of laughs per unit time “2”, age “65”, and gender “female” as keys, and the corresponding “health information” is obtained. I do.
[0106]
In this embodiment, the health information data 1602 is composed of gender, age, average number of laughs per unit time, and health information. However, blood pressure value, cholesterol value, other information, etc. Needless to say, a configuration including The data structure of the health information does not matter.
[0107]
Such health information may be, for example, "Normal number of laughs. It seems that there is no problem with your health condition. Care should be taken to laugh enough." If this condition continues, it may be unfavorable for mental stability and may lead to mental disorders such as depression, dementia, amnesia, etc. In addition, there is a decline in physical functions caused by mental factors and lifestyle-related diseases If you don't laugh, your immune function may be reduced and you may develop an immune disease, cancer, or other illness. Be careful to laugh enough. " And so on.
[0108]
Although the average number of laughs per unit time per month was used as the search condition, the average number of laughs per unit time per unit time during other fixed periods, for example, the average for two weeks or the average for one year But it's fine. Alternatively, health information may be obtained by searching using a change amount of the average number of laughs per unit time for each fixed period (for example, every single month, every two months) as a key. In this case, the data structure of the health information includes data that can be searched by the amount of change in the number of laughs per unit time.
[0109]
In the present embodiment, the information processing apparatus has the health information stored in advance in the health information storage means. However, the information processing apparatus has a means for customizing information about health, The health information stored in the storage means may be updated to the latest health information.
[0110]
Further, the health information acquisition means retrieves the health information stored in advance in the health information storage means and acquires the corresponding health information. However, the health information acquisition means is provided on a server different from the information processing apparatus. The health information may be acquired from a database owned by the company. It goes without saying that a plurality of such servers may be used.
[0111]
Next, the health information output unit acquires the health information configured by the health information configuration unit, and transmits the health information to the health information display device identified by the user identifier. That is, the health information output unit transmits the health information to the device having the IP address corresponding to the identification number included in the user identifier. After the transmission, the operation of the information processing device ends.
[0112]
Next, the health information display device receives the health information and outputs the health information to the liquid crystal display. FIG. 17 shows a display example of health information. According to FIG. 17, the identification number, name, age, gender, average number of laughs per unit time per month, judgment, and health information are displayed. The judgment is information included in the health information. If the number of laughs is equal to or more than a certain value, the state of good health is evaluated. In this embodiment, when the average number of laughs per unit time in a single month is 0, the evaluation is E. When the number is 1 to 2, the evaluation is D, and the number is 3 to 3. In the case of 4, the evaluation is C, when the number is 5 to 6, the evaluation is B, and when the number is 7 or more, the evaluation is A. The fixed value may be another value according to the health information or the like.
[0113]
In the present embodiment, the health information display device outputs the health information to the liquid crystal display. However, the health information display device may store the health information in a recording medium built in the device. Further, the laughter detection device may receive the health information and output the health information. In such a case, the laughter detection device includes a receiving unit that receives the health information and an output unit that outputs the health information.
[0114]
Further, the information processing device may transmit health information in response to a request from a user.
[0115]
In the present embodiment, the number of laughs is detected by the laugh count detecting means of the first embodiment. However, the number of laughs of the user himself is detected by writing down the number of laughs in a memo or the like. Alternatively, the sound may be recorded, and laughter may be detected by other detecting means such as a means for detecting the number of laughs from the recorded sound.
[0116]
Further, in the present embodiment, the health information output unit transmits the health information to the health information display device, but the health information output unit has a liquid crystal display or the like, and the health information configuration unit configures the health information output unit. It may output information. FIG. 18 shows a display example of the health information in such a case. According to FIG. 18, the identification number, name, age, gender, average number of laughs per unit time per month, judgment, and health information are displayed. The age and gender are pre-recorded in a memory built in the information processing device, and the health information output unit searches using the identification number as a key, acquires the age and gender, and outputs the acquired information. is there.
[0117]
Also, in the present embodiment, the health information output unit, the health information reception unit, and the information output unit transmit, receive, and output the health information, respectively. It may transmit, receive, and output.
[0118]
Further, in the present embodiment, the judgment output unit, the laughter information receiving unit, the health information output unit, and the health information receiving unit transmit or receive respective information via the Internet. The information may be transmitted or received by such means.
[0119]
Further, the determination result output unit and the health information output unit may be configured by broadcasting means and the like, and may broadcast each information, and the laughter information receiving unit and the health information receiving unit may receive the respective information.
[0120]
As described above, the information processing apparatus according to the present embodiment configures health information relating to mental health based on the number of laughs detected by the laugh detection device and transmits the health information, and determines the number of laughs as a criterion. As a result, it is possible to easily obtain information on a mental health condition. Such health information can be used as one of the determination indices of QOL (Quality of Life). Further, detailed information can be obtained on the adverse effects caused by not laughing, and the user can pay sufficient attention to laughing and maintain a good health condition. Also,
[0121]
Further, a program when the operation of the laughter detection device in the present embodiment is realized by software is as follows. The computer has a sound acquisition step of acquiring sound, a signal extraction step of extracting only a signal in a certain frequency band from the sound acquired by the sound acquisition step, and a sound acquisition step based on the signal extracted by the signal extraction step. This is a program for executing a determination step of determining whether the acquired sound is laughter, and a determination result output step of outputting laughter information including a determination result and a user identifier in the determination step.
[0122]
Further, a program when the operation of the information processing apparatus according to the present embodiment is realized by software is as follows. A program for causing a computer to execute a laughter information receiving step of receiving laughter information, a health information configuration step of configuring health information based on the laughter information, and a health information output step of outputting health information. .
【The invention's effect】
According to the present invention, an accurate number of laughs can be detected, and health information on mental health can be easily acquired based on the number of laughs.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of a laughter detection device according to a first embodiment.
FIG. 2 is a flowchart illustrating an operation of the laughter detection device according to the first embodiment;
FIG. 3 is a flowchart illustrating an operation of the laughter detection device according to the first embodiment;
FIG. 4 is a diagram showing a bone conduction microphone and a bone conduction microphone when worn;
FIG. 5 is a diagram showing a laughter detection device.
FIG. 6 is a diagram showing an example of an audio signal for each frequency band.
FIG. 7 is a diagram showing a time-series change of amplitude envelope data of a vowel frequency band signal.
FIG. 8 is a diagram showing experimental results obtained by measuring a period of laughter and a period of conversation.
FIG. 9 is a diagram showing variations in the cycle of laughter and conversation.
FIG. 10 is a diagram showing a display example of the number of laughs output on the liquid crystal display.
FIG. 11 is a block diagram showing an information processing system according to a second embodiment.
FIG. 12 is a block diagram illustrating a configuration of a health information configuration unit of the information processing apparatus according to the second embodiment.
FIG. 13 is a diagram showing an appearance of an information processing system according to a second embodiment.
FIG. 14 is a diagram showing an example of laughter information;
FIG. 15 is a diagram showing an example of searching for age and gender.
FIG. 16 is a diagram showing an example of searching for health information.
FIG. 17 shows a display example of health information.
FIG. 18 is a diagram showing a display example of health information.
[Explanation of symbols]
101, 503 Start reception unit
102, 401, 402, 502 sound acquisition unit
103, 504 Judgment start reception unit
104 signal extraction unit
105 Judgment unit
106, 505, 1111 judgment result output unit
501, 1110, 1302 laughter detection device
1120, 1301 Information processing device
1121 laughter information receiver
1122, 1101 and 1210 Health information component
1123 Health information output unit
1130, 1303 Health information display device
1131 Health information receiving unit
1132 Information output unit
1211 Health information storage means
1212 Health information acquisition means
1501 Examples of data such as name, age, and gender
1601 Examples of average number of laughs, age, and gender per unit time with search key
1602 Example of health information data
1603 Example of recording medium with accumulated health information

Claims (12)

A sound acquisition unit for acquiring sound;
A signal extraction unit that extracts only a signal in a certain frequency band from the sound acquired by the sound acquisition unit,
Based on the signal extracted by the signal extraction unit, a determination unit that determines whether the sound acquired by the sound acquisition unit is laughter,
A laughter detection device comprising a judgment result output unit for outputting a judgment result in the judgment unit. The laughter detection device according to claim 1, wherein the sound acquisition unit acquires sound using a bone conduction microphone arranged on a larynx. 3. The laughter detection device according to claim 1, wherein the signal extraction unit extracts only a specific frequency band signal from the wavelet transform. 4. 4. The laughter detection device according to claim 1, wherein the determination unit determines a period in which the amplitude peak is equal to or more than a predetermined value and which periodically appears in the signals extracted by the signal extraction unit as a laughter period. 5. apparatus. 5. The laughter detection device according to claim 4, wherein the period is about 80 ms to about 120 ms. The laughter detection device according to any one of claims 1 to 5, wherein the determination result output unit outputs the number of times that the determination unit determines that the laugh is performed. A laughter information receiving unit that receives laughter information having laugh information that is information about laughter,
Based on the laughter information, a health information configuration unit that configures health information that is information about mental health,
An information processing apparatus comprising a health information output unit that outputs the health information. The laughter information includes information on the number of laughs per unit time, and the health information forming unit, if the information on the number of laughs per unit time is information indicating the number of laughs equal to or less than a certain number of times. 8. The information processing apparatus according to claim 7, wherein information indicating that the subject is unhealthy is configured as health information. The laughter detection device, an information processing system including the information processing device,
The laughter detection device,
Detecting the number of laughter, transmitting laughter information having laugh information that is information about laughter to the information processing device,
The information processing device,
Receiving laughter information having laugh information that is information about laughter,
Based on the laughter information, configure health information that is information about mental health,
An information processing system for outputting the health information; The laughter detection device, an information processing system including the information processing device,
The laughter detection device,
Detecting the number of laughter, transmitting laughter information having laugh information that is information about laughter to the information processing device,
The information processing device,
Receiving laughter information having laugh information that is information about laughter,
The information processing system according to claim 9, wherein health information that is information on mental health is transmitted to a predetermined device based on the laughter information. A sound acquisition step of acquiring sound;
A signal extraction step of extracting only a signal in a certain frequency band from the sound acquired in the sound acquisition step,
Based on the signal extracted in the signal extraction step, a determination step of determining whether the sound acquired in the sound acquisition step is a laugh,
A laughter detection method comprising a judgment result output step of outputting a judgment result in the judgment step. On the computer,
A sound acquisition step of acquiring sound;
A signal extraction step of extracting only a signal in a certain frequency band from the sound acquired in the sound acquisition step,
Based on the signal extracted in the signal extraction step, a determination step of determining whether the sound acquired in the sound acquisition step is a laugh,
A program for executing a judgment result output step of outputting a judgment result in the judgment step.

Images