JP6863563B2 - Stress evaluation system - Google Patents

Description

The present invention relates to a method and a system for evaluating a person's psychological stress. In particular, regarding stress evaluation methods and stress evaluation systems that enable selective and continuous evaluation of bad stress (Distress), which is said to be highly likely to cause depression, from among various types of stress. ..

Recently, it has become a social problem that mental damage caused by the daily continuation of bad stress and fatigue has accumulated, and if this condition continues, the number of cases of depression is increasing. In today's busy society, it is inevitable that stress and fatigue will continue frequently. However, it is also known that moderate stress works as good stress (Eustress) for humans. In other words, the existence of stress is unavoidable, but the way individuals receive it and the degree of accumulation vary widely, and it is necessary to know what kind of stress an individual is in every day and to adjust to this condition. It is important to take a rest or change the work contents accordingly.

Stress is said to be a general term for non-specific reactions that occur in the human body in response to external stimuli to humans (Non-Patent Document 1). And a person's response to stress varies greatly from person to person. When a person is stressed, he or she first causes functional deterioration (shock phase), but then reverses and increases function to cope with stress (anti-shock phase). In this state, a person can be said to be good stress (Eustress) if it has a positive effect as a result, and conversely it can be said to be bad stress (Distress) if it has a negative effect on the person. .. It is known that if a state of strong distress continues, it causes depression (Non-Patent Document 2).

Conventionally, as described above, the measurement of stress having individual differences is a method of evaluating stress by a physiological index by measuring brain waves, skin surface temperature, blood flow, skin electrical activity, electrocardiogram, etc., and subjective evaluation. As a method of obtaining a psychological index, means such as interviews, questionnaires, etc. have been performed so as to meet various purposes. For example, at the time of health examination in the workplace, means such as a doctor's interview, a questionnaire, etc. have been usually used. Moreover, in hospitals, in addition to means such as interviews with doctors, questionnaires, etc., a method of evaluation by measuring the above physiological indicators is used for stress measurement of depressed patients who are considered to be excessively stressed. I came.

For example, according to Patent Document 1, the stress level is calculated from the waveform of the pulse wave. Attempts have also been made to Lorentz plot the variation in the peak value interval (R-R interval) of the pulse wave and measure the stress level from the distribution at this time (Non-Patent Document 3). This method is non-invasive and can be measured by a simple method, and is excellent as a means for daily measurement. However, with this means, although it can be determined from the distribution that stress is applied, it is difficult to determine whether the stress is good stress (Eustress) or bad stress (Distress).

As a method for evaluating stress by other physiological indexes, an attempt by attaching a nerve sensor to the scalp and measuring brain waves (Patent Document 2), an attempt by measuring sweating (Patent Document 3), and an attempt by measuring the amount of amylase activity in saliva (Patent Document 4), etc., there are many attempts. These measurements have problems that special equipment is required for the measurement, a long time is required to obtain the result, and the stress is not suitable for continuous evaluation on a daily basis. Moreover, in each method, it was difficult to determine whether the stress was good stress (Eustress) or bad stress (Distress).

On the other hand, conventionally, the judgment of whether it is good stress (Eustress) or bad stress (Distress) is mainly based on the observation by a doctor by interviewing the person to be measured and the subjective evaluation by a questionnaire survey of the person to be measured. It was mainstream.

However, with this method, it is difficult to say that an objective judgment is made because it largely depends on the subjectivity of the doctor or psychologist regarding the mental aspect of the subject, or the subjectivity and declaration of the subject himself. .. In addition, since diagnosis takes time, it is not suitable for stress measurement methods and systems for continuous and simple measurement. And there was a high possibility that the change and deterioration of the mental status of the subject to be measured would be overlooked.

Then, in recent years, in order to objectively diagnose this symptom of a patient who has become depressed as a result of a continuous strong state of bad stress (Distress), an attempt by measuring components in the body fluid of the patient (Patent Documents). Attempts have been made to use 5) and blood components as a guideline for the diagnosis of depression (Patent Document 6).

However, these measurements are unsuitable for stress evaluation methods and systems for continuous and easy measurement because the measuring device is complicated, expensive, and time-consuming. And there was a high possibility that the change and deterioration of the mental status of the subject to be measured would be overlooked.

Japanese Patent No. 330070346 Special Table 2014-527219 JP-A-2007-283095 Japanese Unexamined Patent Publication No. 2007-275287 Japanese Unexamined Patent Publication No. 2014-236479 International Publication No. 2006-105907

Selye HA. J Clin Endocrinnol 1946; 6; 117-230. "Science and Health of Stress" edited by Takeo Futaki, p2-p7, published in January 2008, Kyoritsu Shuppan Matsumoto et al., Research on mental stress evaluation method due to heart rate fluctuation, Life Support Vol.22, No.3, 2010

The problem to be solved is that the stress level can be measured by the method of evaluating the stress by the physiological index to evaluate the stress state of the person to be measured, but it is possible to grasp whether it is bad stress or good stress. There are problems such as difficulty, special equipment is required for measurement, it takes a long time to obtain results, and it is not suitable for continuously evaluating stress on a daily basis. In addition, the method of obtaining a psychological index, which is a subjective evaluation, can grasp whether it is bad stress or good stress, but it does not make an objective judgment, it takes time to diagnose, and stress that is continuously and easily measured. There were problems such as unsuitability for the measurement method and system, the ever-changing changes in the mental status of the subject, and the high possibility of overlooking the deterioration.

The present invention has been made to solve such a problem, and the purpose of the present invention is to grasp whether it is bad stress or good stress, to evaluate it in a short time, and to require an expensive device. It provides a quantitative, objective, stress evaluation method and stress evaluation system that does not depend on the subjectivity of the doctor and does not require the entry of the subject.

In order to solve the above problems, the stress evaluation method and the stress evaluation system according to the present invention provide the following.

(1) When the person to be measured is in a healthy state, the person to be measured is photographed with a camera before and after the stress is applied, and the feature points are taken from the image data of the body of the person to be measured. Is extracted, the coordinates of these feature points are stored in the standard database, and the difference between the coordinates of the feature points when the subject is in a healthy state and the coordinates of the feature points at the time of measurement stored in the standard database is vectorized. A stress evaluation method that evaluates the stress that the person being measured receives at the time of measurement from this vector.

According to the present invention, it is possible to grasp whether the stress is bad or good from the vectorization of the feature points extracted from the body image data of the photographed subject and the standard body feature points, and the measurement and the evaluation can be performed in a short time. It is possible, and stress can be assessed without the need for expensive equipment to measure, without the subjectivity of the doctor, and without the need for subject entry. In addition, by accumulating the vector data of the extracted feature points each time the measurement is performed, the number of samples of the data in the standard state of the person to be measured increases, and by increasing the number of samples of the data in the standard state, the data accuracy of the standard state increases. It is possible to more accurately evaluate the stress state of the person to be measured at the time of measurement.

(2) The stress evaluation method according to (1), wherein the face of the person to be measured is photographed with a camera, and the feature points of the facial expression of the person to be measured are extracted from the photographed face image data.

According to the present invention, the face of the person to be measured is photographed with a camera, and the facial image feature points of the person to be measured are vectorized from the captured face image data and the facial image feature points in a healthy state. It is possible to grasp whether it is stress, measurement and evaluation can be done in a short time, measurement does not require expensive equipment, does not depend on the subjectivity of the doctor, and evaluates stress without the need to fill in the subject. Can be done.

(3) The stress evaluation method according to (1), wherein the whole body of the person to be measured is photographed with a camera, and the characteristic points of the person to be measured are extracted from the captured moving image data.
The stress evaluation method according to (1), wherein the face of the person to be measured is photographed with a camera, and the feature points of the facial expression of the person to be measured are extracted from the photographed face image data.

According to the present invention, it is possible to grasp whether the stress is bad or good from the vectorization of the gesture feature point movement of the person to be measured and the standard gesture feature point movement extracted from the photographed gesture image data. The measurement and evaluation can be performed in a short time, and the stress can be evaluated without requiring expensive equipment for measurement, independent of the subjectivity of the doctor, and no entry of the subject.

(4) There are a plurality of people to be measured, and the standard database stores the coordinate data of the feature points of the plurality of people to be measured. The stress evaluation method according to (1), (2), and (3), which includes a step of obtaining a vector based on the difference between the coordinates of the feature point in a healthy state and the coordinates of the feature point in measurement. ..

According to the present invention, in order to evaluate the stress received by the person to be measured, a vector based on the difference between the coordinates of the feature points in a healthy state other than the person to be measured and the coordinates of the feature points at the time of measurement is also used. Since the steps to be obtained are included, the personal characteristics of the person to be measured can be extracted, and by increasing the number of samples of data in the standard state, the data accuracy in the standard state can be improved, and thus the data accuracy at the time of measurement can be improved. The stress state of the person to be measured can be evaluated more accurately.

(5) The standard database consists of a standard database created from the entire subject, a standard database created from a specific group, and a personal standard database created from the past measurement results of the individual subject. The stress evaluation method according to (4).

According to the present invention, each time the movement data and the heartbeat data of the extracted feature points are measured, the data as an individual, the data as a group, and the data as a whole are accumulated as the data of the individual to be measured. The number of samples in the standard state, the standard state as a group, and the standard state as a whole has increased, the data accuracy of the standard state has been improved with each measurement, and the stress state of the person to be measured at the time of measurement has become more accurate. Can be evaluated.

(6) The emotion of the person to be measured is based on the feedback consisting of the step of identifying the frequent pattern of the vector from the vector obtained from the difference in coordinates, the result of the questionnaire from the person being measured, the result of the interview, and the result of the person's declaration. The stress evaluation method according to (4) and (5), which includes a step of obtaining a correlation value of.

According to the present invention, the relationship between the frequent pattern of the vector of the person to be measured and the stress state of the person to be measured is specified, and the result of a questionnaire, the result of an interview, or the result of the person's declaration regarding the stress from the person to be measured is used. It is possible to obtain the relationship with the feature point vector at the time of bad stress and to specify the stress state at the time of measurement of the person to be measured. Thereby, the stress evaluation method according to the present invention can be performed in a short time, and the stress can be evaluated without requiring an expensive device for measurement, independent of the subjectivity of a doctor, and no entry of a subject. You can.

(7) The stress evaluation method according to (1)-(6), wherein the step of measuring the pulse wave of the person to be measured is included.

According to the present invention, it is possible to grasp whether the stress is bad or good from the facial image feature vector extracted from the moving image data of the face of the person to be measured taken by the camera, and the pulse wave of the person to be measured. The stress can be quantitatively determined from the result of the measurement, and both measurement and evaluation can be performed in a short time, the measurement does not require an expensive device, does not depend on the subjectivity of the doctor, and the subject's subject. Stress can be assessed without the need for entry.

(8) The stress according to (1)-(7), which includes a step of obtaining a Lorentz plot of a change in the interval of the measured peak value of the heartbeat and a step of evaluating the degree of stress from the distribution of the Lorentz plot. Evaluation method.

According to the present invention, it is possible to grasp whether the stress is bad or good from the extracted facial image feature vector, and the peak value interval (RR interval) of the pulse wave is obtained from the result of measuring the pulse wave of the person to be measured. The distribution of the Lorenz plot, which plots the distribution of changes in, is obtained, and the degree of stress is quantitatively evaluated from the results, the stress can be quantitatively obtained, and both measurements and evaluations can be performed in a short time. There is no need for expensive equipment for measurement, it does not depend on the subjectivity of the doctor, and stress can be evaluated without requiring the entry of the subject.

(9) The stress evaluation method according to claim 1-8, wherein the talk of the person to be measured is recorded by a microphone, feature points are extracted, and stress is evaluated in addition to the feature points.

According to the present invention, the characteristic is vectorized by comparing the measurement-time chatter of the person to be measured and the chatter of the standard chatter feature database recorded by the microphone, and the facial expression feature vector data and the heartbeat data are combined to receive the subject. It is possible to evaluate the stress of the measuring person during measurement. By these measurements, it is possible to grasp whether the stress is bad or good, the stress can be quantitatively obtained from the result of measuring the heartbeat of the person to be measured, and both measurement and evaluation are short-time. It is possible to evaluate stress without the need for expensive equipment for measurement, independent of the subjectivity of the doctor, and no entry of the subject.

(10) The camera that measures the person to be measured, the recording unit that records the moving image taken by the camera, the feature point extracting unit that extracts the feature points from the recorded moving image data, and the extracted feature points are compared. A feature point comparison unit, a data accumulation unit that accumulates the extracted data, a data comparison unit that compares the accumulated data with the data measured this time, and a judgment unit that judges stress from the compared data. Includes a pulse wave measurement unit that measures pulse waves, a Lorenz plot generation unit that obtains a Lorentz plot of changes in the interval of peak values of measured pulse waves, and a stress evaluation unit that evaluates the degree of stress from the distribution of Lorentz plots. A stress assessment system characterized by this.

According to the stress evaluation system according to the present invention, it is possible to grasp whether the stress is bad or good from the movement data of the feature points extracted from the captured video data, and the stress is measured from the result of measuring the heartbeat of the person to be measured. Can be quantitatively determined, and both measurements and evaluations can be performed in a short time, the measurement does not require expensive equipment, does not depend on the subjectivity of the doctor, and does not require the entry of the subject. Stress can be evaluated.

(11) The pulse wave measuring unit calculates a pulse wave from the color change measuring unit that measures the change in skin color due to the blood flow of the subject's face measured by the camera and the measured color change. The stress evaluation system according to (10), which comprises a pulse wave calculation unit.

According to the stress evaluation system according to the present invention, it is possible to grasp whether the stress is bad or good from the movement data of the feature points extracted from the video data taken by the camera, and to measure the heartbeat of the person to be measured. Therefore, the stress can be calculated quantitatively, both measurement and evaluation can be performed in a short time, the measurement does not require an expensive device, does not depend on the subjectivity of the doctor, and requires the entry of the subject. You can evaluate stress without doing it.

(12) The stress evaluation system according to (10) and (11), wherein the camera is composed of a plurality of cameras.

According to the present invention, it is possible to grasp whether the stress is bad or good from the movement data of the feature points extracted from the video data taken stereoscopically by a plurality of cameras, and the heartbeat of the person to be measured. The stress can be quantitatively obtained from the result of the measurement, and both measurement and evaluation can be performed in a short time, the measurement does not require an expensive device, does not depend on the subjectivity of the doctor, and the subject's subject. Stress can be assessed without the need for entry.

(13) The stress evaluation system according to claims 10 and 11, wherein the talk of the person to be measured is recorded by a microphone, feature points are extracted, and the feature points are used for stress evaluation.

According to the present invention, in addition to the captured video data, the characteristics related to the talk of the person to be measured by the microphone can be measured at the same time, and the stress is quantitatively obtained from the result of measuring the pulse wave of the person to be measured. It is possible, and both measurements and evaluations are possible in a short time, and the measurement does not require expensive equipment, does not depend on the subjectivity of the doctor, and does not require the subject's entry to evaluate the stress. Can be done.

(14) The stress evaluation system according to (10), wherein the camera is attached to a gate through which the person to be measured passes, and photographs the person to be measured in conjunction with the passage of the person to be measured through the gate.

According to the present invention, since the camera is attached to the gate through which the person to be measured passes, it is possible to acquire data every time the person to be measured passes through this gate, and the stress of the person to be measured changes daily. The condition can be evaluated.

(15) The camera is attached to a room where the person to be measured conducts a health examination, a meeting, a meeting, work or work, and the person to be measured participates in a health examination, a meeting, participates in a meeting, performs work or performs work. The stress evaluation system according to (10), wherein the person to be measured is photographed in conjunction with the above.

According to the present invention, the stress of the person to be measured changes daily because the camera photographs the person to be measured in conjunction with the person to be measured to take a picture of the person to be measured in conjunction with the medical examination, participation in a meeting, participation in a meeting, business execution or work execution. The condition can be evaluated.

From short-time moving images or still images taken from a camera, it is possible to extract facial expressions and body movement features that are optimal for determining whether stress is bad or good.
With an inexpensive device, it is possible to measure the stress leading to depression on a daily basis by a simple method.

In the figure showing the basic flow of the stress evaluation method according to the present invention, 10 is the start, 11 of the flow on the left side of the figure is the image acquisition of the person to be measured, 12 is the feature point extraction, 13 is the comparison with the DB, and 14 is the feature. Point vector calculation, 15 is stress analysis, 16 is good / bad stress judgment, 17 is database, 18 of the flow on the right side of the figure is pulse wave acquisition, 19 is RR value calculation, 20 is Lorentz plot calculation, 21 indicates Lorentz plot area calculation, 22 indicates area comparison with DB, 23 indicates stress level calculation, 24 user opinions are input to both results, 25 is stress evaluation, and 26 adds the result to DB data. , 27 indicates the quality of the stress evaluation result, 28 indicates the alert, and 29 indicates the report. It is a figure showing the feature points extracted from the face image of the person to be measured taken by the camera. 31 is eyebrows, 32 is eye contour, 33 is eye center, 34 is nose contour, 35 is upper lip contour, 36. Indicates the contour of the upper lip, and 37 indicates the contour of the face. The contents of each database are shown in the database data structure: (a) before stress is applied, (b) after good stress is applied, and (c) after bad stress is applied. The characteristic points when stress is applied to the person to be measured are (a) angry characteristic points and (b) sadness characteristic points, respectively. In the acquired heart rate information, 51 indicates an Rn peak and 52 indicates an Rn + 1 peak. The Lorentz plot shows (a) before stress is applied and (b) after stress is applied. In the block diagram of the first embodiment of the stress evaluation system, 71 is a camera, 72 is a display, 73 is a PC unit, 74 is a database, 75 is an internet line, and 76 is a heart rate measuring device. In the figure illustrating the second embodiment of the stress evaluation system, the feature points of the face are extracted and the heartbeat is measured from the skin area at the same time by the camera. 81 indicates a facial feature point, and 82 indicates a skin area.

FIG. 1 is a flowchart illustrating the best mode for carrying out the stress evaluation method according to claims 1, 2, 4, 5, 6, 7, and 8.

As shown in FIG. 1, in the stress evaluation method according to the present embodiment, a flow of taking a face image of a person to be measured as a flow on the left side to determine whether it is bad stress or a good stress, and a flow to be measured as a flow on the right side. It consists of a flow for measuring a person's pulse wave and calculating a stress level, and a flow for evaluating the stress of a person to be measured using the results of both.

The first step in the flow on the left side is the step of acquiring the image of the person to be measured 11 in which the face image of the person to be measured is photographed as a stereoscopic still image by two cameras and the photographed image is acquired in the recording unit of the PC. .. In this embodiment, a color image was taken with a camera having 1980 × 1080 pixels. Depending on the purpose of evaluation by the person to be measured, it is possible to select monochrome or color, or a high resolution of 1980 x 1280 from a low resolution of about 640 x 480 pixels. In addition, when the number of cameras is one, a two-dimensional image can be obtained, and when a plurality of cameras are used, a three-dimensional image of the face can be obtained. If you want to capture the momentary movement of blinks, mouth, and cheeks, take a moving image.

The second step in the flow of measuring and evaluating the face image on the left side is to extract the feature points from the face image of the person to be measured saved in the first step and specify the coordinates as shown in FIG. There are 22 steps. In this embodiment, 78 feature points were extracted. 16 points along the contour 31 of the left and right eyebrows, 16 points along the contour 32 of the eye, 2 points along the center 33 of the eye, 17 points along the contour 37 of the face, and 7 points along the contour 34 of the nose. , 10 points were distributed along the contour 35 of the upper lip, and 10 points were distributed along the contour 36 of the lower lip. In this distribution, the influence of the mental state of the person to be measured easily reflects the position and shape of the eyebrows, eyes, eye movements, nose, mouth, etc. with respect to the contour of the face, and the facial expression changes at 78 feature points. This is because it is possible to grasp.

In this embodiment, feature points are set for the reasons described above, but the present invention is not limited to this. It is also possible to set in more detail and select the part of the face to pay attention to. In another embodiment, this score was set to 1000 points. In this case, the feature points can be extracted more accurately, but the amount of calculation becomes enormous, and the comparison process with the database takes a lot of time. This score can be selected according to the purpose of evaluation and the ability of the system.

To specify the coordinates of the feature points, the apex of the nose in the center of the face is selected as the origin, and the direction and scale of the coordinate axes are determined according to the positions of the feature points on the contour of the face. However, since these coordinates are relative coordinates, the position of the origin, the axial direction, and the scale are not limited to this selection method.

The third step in the flow of measuring and evaluating the face image on the left side is to compare the coordinates of the feature points of the face at the time of measurement of the person to be measured with the coordinates of the feature points stored in the database and express the difference as a vector. There are 13 steps.

The structure of the database (DB) will be described with reference to FIG. The database contains data for all targeted subjects. The data of each person to be measured is stored in three states: (a) before stress is applied (state A), (b) after stress is applied (state B), and (c) after bad stress is applied. Has been done.

The structure of each data is evidenced by the results of a questionnaire conducted on the subject, the results of the diagnosis of the subject by a doctor, the PPT test for stress tolerance, and the measurement of the peak interval (RR interval) by heart rate measurement. The correlation coefficient between the result of the stress state of the measuring person and the facial feature point vector of the measuring person is scored.

Next, in the fourth step in the flow of measuring and evaluating the face image on the left side, the coordinate vector calculation 14 of the feature points in FIG. 1 is performed. The coordinate vector of the feature point (hereinafter, simply referred to as a vector) is determined by the coordinates of the face feature point of the person to be measured at the time of measurement and the coordinate difference and direction before stress is applied to the person to be measured (state A) in the database. For example, FIG. 4 (a) shows the facial feature points when the subject is stressed and angry, and FIG. 4 (b) shows the facial feature points when the subject is stressed and sad. It is a figure.

In this embodiment, 78 feature point vectors are determined. For example, when there are wrinkles between the eyebrows, the feature points on both sides of the eyebrows are upward vectors and the inner feature points are lowered, or the vector corresponding to the shape of the mouth is large. Or, it is a vector that corresponds to the surprised opening of the eyes, a vector of eye movements that do not settle down, or a vector movement that the eyelids frequently open and close. Etc., the facial expression of the person to be measured at that time is calculated by a vector.

Next, the stress analysis 15 is performed in the fifth step in the flow of measuring and evaluating the face image on the left side. In this analysis, the vector value at the time of measurement of the person to be measured is compared with the vector at the time of stress load in the database, and the correlation value is obtained to see how much the vector at the time of measurement corresponds to the combination of the vectors at the time of stress load. Then, it is determined that the correlation value corresponds to the stress state having the largest value.

Although it is possible to target all values in the method of obtaining the correlation, in order to perform the calculation efficiently, the stress is specified in advance for each frequent vector in the vector value in the DB, and the stress is measured. Stress is efficiently identified by finding which frequent vector the time vector value corresponds to.

When performing the above stress analysis, the area corresponding to the distribution formed by the points of the Lorentz plot of the person to be measured for the pulse wave obtained in the flow on the right side of FIG. 1 (how to obtain this area will be described later) is also referred to. ..

Then, in the sixth step in the flow of measuring and evaluating the face image on the left side, whether or not the stress of the state at the time of measurement of the person to be measured is applied, and if the stress is applied, the stress is good. A good / bad stress determination 16 for determining whether the stress is stress or bad stress is performed. In this determination, the feature point vector when the feature point vector at the time of measurement of the person to be measured is closest to the vector of the corresponding feature point in the DB is obtained, and the correlation with the stress state with evidence in the DB at this time is obtained. Judgment is made based on a few scores.

In the sixth step, what is judged to be bad stress is the doctor's diagnosis result, the questionnaire to the person to be measured, and the facial feature point vector that frequently appears in the case of bad stress from the PPT test and RR interval fluctuation. When the correlation coefficient score with the feature point vector of the face at the time of measurement is higher than the reference value, it is judged as bad stress. On the contrary, what is judged to be good stress is the diagnosis result of the doctor, the questionnaire to the person to be measured, the facial feature point vector that frequently appears in the case of already good stress from the PPT test and the fluctuation of the RR interval, and the measured person. When the correlation coefficient score with the feature point vector of the face at the time is higher than the reference value, it is judged as good stress.

Next, the flow on the right side of FIG. 1 for measuring and evaluating the heartbeat will be described.

The first step on the right side of FIG. 1, the heartbeat pulse wave acquisition 18, is continuously acquired by the heartbeat measuring device for about several minutes. There are various types of heart rate measuring devices, such as a type that takes a pulse wave with a finger and a type that takes a pulse wave from the chest near the heart, depending on the purpose, and the heart rate measuring device is selected according to the purpose. In this embodiment, since it is sufficient that the interval between the peak values of the pulse wave (RR interval) can be measured, a device capable of measuring the pulse wave from a simple finger was used. FIG. 5 shows the pulse wave measurement result, where 51 is Rn indicating the nth peak and 52 is Rn + 1st peak indicating the n + 1th peak.

In the second step on the right side of FIG. 1, the interval between the peak values of the pulse waves acquired in the first step is calculated. As shown in FIG. 5, the acquired pulse wave data is repeated with peak values, and the interval between the peak values (RR interval) is calculated.

Next, in the third step on the right side of FIG. 1, the R-R interval is taken as the nth on the X-axis and the n + 1th on the Y-axis for each pulse, and plotted on the XY plane to calculate the Lorentz plot. To do. Fig. 6 (a) is a Lorentz plot when the subject is not stressed. When the subject is in a stable and relaxed state, the RR interval fluctuates and the plot points spread to both Y = X straight lines. .. On the other hand, FIG. 6 (b) is a Lorentz plot when the subject is stressed. In a state where the mental tension is caused by the stress, the fluctuation of the RR interval decreases, and the plot points become a Y = X straight line. On the other hand, the spread is reduced.

In the fourth step on the right side of FIG. 1, the area corresponding to the distribution formed by the points of the Lorentz plot of the subject is the contour of the distribution formed by the standard deviation of the distance of each point from the Y = X straight line. Is obtained, and the Lorentz plot area calculation 21 is performed in which the area formed by this contour is used as the Lorentz plot area.

Next, in the fifth step on the right side of FIG. 1, the obtained Lorentz plot area, the Lorentz plot area of the person to be measured in the DB, the Lorenz plot area of the group to which the person to be measured belongs, and the Lorenz plot of the entire person to be measured The area is compared with the area 22 respectively.

Next, in the sixth step on the right side of FIG. 1, how much the stress level of the person to be measured is compared with the data obtained so far, and the person to be measured. Compared with the stress level of the members of the same group as the above, and compared with the stress level of all the members to be measured, whether there is a significant difference in the stress level with respect to the other members, the equal stress level calculation 23 is performed. To.

The stress evaluation 25 is performed based on the result of the good / bad stress determination 16 obtained from the face image in the left flow of FIG. 1 and the stress level calculation 23 result obtained from the pulse wave in the right flow of FIG. In the present embodiment, if both data and the opinion of the person to be measured are heard, the result is input 24, and here, whether the stress is good or bad and how intense the stress level is. Results such as how much the stress state of the person to be measured is in the group to which the person to be measured belongs, how much is the stress state in the whole person to be measured, and the like can be obtained.

Then, the evaluated stress evaluation result is added to the database as data26. Each time such a measurement is made, the measurement data increases and the accuracy of the standard data stored in the database increases.

Finally, the alert 28 is issued for these stress evaluation results according to a predetermined standard, and the result report 29 is output. In this embodiment, an alert is issued when it is determined that bad stress is applied, and a report is output when it is determined that good stress is applied.

Effect of this example

In this embodiment, only a camera and a pulse wave measuring device are used for measurement, and the measurement results are also processed by a DB stored in a normal personal computer and a server, and measurement and evaluation can be performed on a daily basis with an inexpensive device. In addition, the measurement and evaluation were performed in a short time, it was possible to determine whether the stress was bad or good, and the degree of stress applied to the person to be measured could be quantitatively evaluated.

In the above embodiment, the evaluation was performed using a still image of the face, but in the invention of claim 3, the whole body of the person to be measured is photographed with a camera, the movement of the feature points of the whole body is extracted, and the whole body is evaluated. By comparing the vector of the feature points of the above with the standard database and following the same steps as in the above embodiment, it is determined whether the stress is bad stress or good stress.

For example, the motion vector is evaluated, such as the roundness of the back is increased from the standard time, the movement of the limbs is awkward than the standard time, and the like.

In the above embodiment, the evaluation was performed using an image of the face or body, but further, the talk of the person to be measured is recorded with a microphone, this feature point is extracted, and the same steps as in the above embodiment are followed. By doing so, it is judged whether the stress is bad stress or good stress.

For example, a motion vector such as a more muffled talk or a quiet talk is evaluated.

Next, a first embodiment for implementing the stress evaluation system according to claim 10 will be described with reference to FIG.

As shown in FIG. 7, the stress evaluation system according to the present embodiment data-processes the face of the person to be measured, the camera 71 for photographing the body, the monitor 72 for displaying the captured image, and the captured image. It is composed of a personal computer 73, a database 74 for storing standard data, a net line 75 for connecting the database and the personal computer, and a pulse wave measuring device 76 for measuring the pulse wave of the person to be measured. With this system, stress evaluation can be performed at the same time when employees use Sokon (report material creation, data entry, education report creation) in the company. In addition, at the start of using the personal computer, it is possible to confirm the identity and acquire the image at the same time in the confirmation of access control to the system. In addition, employees usually spend a lot of time using a personal computer, which enables long-term measurement for stress diagnosis.

Next, a second embodiment for implementing the stress evaluation system according to claim 11 will be described. In the first embodiment, the pulse wave of the person to be measured was measured by a heart rate measuring device, but in this embodiment, the pulse wave is calculated from the color signal of the skin portion of the person to be measured measured by the camera 01. .. FIG. 8 is a diagram illustrating extraction of facial feature points 81 and extraction of skin region 82 by a 3D camera. The red component of the skin part changes in synchronization with the pulse wave, and if the pulse wave is measured from the camera, it is possible to acquire the face image and measure the pulse wave from the same camera, and stress with a simpler measurement system. Evaluation becomes possible.

Next, a third embodiment for implementing the stress evaluation system according to claim 12 will be described. In the first embodiment, the face image of the person to be measured was taken with one camera and the feature points were extracted, but the two cameras were arranged at regular intervals with respect to the person to be measured to obtain a stereoscopic image. Can be obtained. With this shooting, the unevenness of the face becomes clearer, and the accuracy of the vector value of the movement of the feature points can be improved.

Next, a fourth embodiment for implementing the stress evaluation system according to claim 13 will be described. In the first embodiment, the face or body image of the person to be measured was taken with a camera to extract the feature points, but in this embodiment, the talk of the person to be measured is recorded by a microphone, the feature points are extracted, and the stress evaluation is performed. It can be used for improving the accuracy of stress evaluation.

Next, a fifth embodiment for implementing the stress evaluation system according to claim 14 will be described. In the first embodiment, the face or body image of the person to be measured was taken with the camera connected to the PC, and the feature points were extracted. In this embodiment, the camera is attached to the gate through which the person to be measured passes. Then, the person to be measured takes a picture of the person to be measured in conjunction with passing through the gate. From the captured image of the person to be measured, the person can be confirmed and the stress of the person to be measured can be evaluated at the same time, and the stress can be measured constantly.

Next, a sixth embodiment for implementing the stress evaluation system according to claim 15 will be described. A camera that captures the person to be measured is attached to the room where the person to be measured conducts a medical examination, a meeting, a meeting, work or work, and the person to be measured participates in a medical examination, a meeting, participates in a meeting, performs work or performs work. If the person to be measured is photographed in conjunction with this, a stress evaluation system that can evaluate stress while performing normal work at the company can be obtained.

The stress evaluation system used in this embodiment does not require a special device for stress evaluation, can be evaluated without interrupting normal work for measurement, and the mental status of the person to be measured changes from moment to moment. It is possible to evaluate without overlooking the change and deterioration of.

10 Start 11 Acquisition of image of person to be measured 12 Extraction of feature points 13 Comparison with DB 14 Calculation of feature point vector 15 Stress analysis 16 Good or bad stress judgment 17 Database (DB)
18 Pulse wave acquisition 19 R-R value calculation 20 Lorenz plot calculation 21 Lorenz plot area calculation 22 Area comparison with DB 23 Stress level calculation 24 User opinion input 25 Stress evaluation 26 Add to DB data 27 Good / bad judgment 28 Alert 29 Report 31 Eyebrows 32 Eye contour 33 Eye center 34 Nose contour 35 Upper lip contour 36 Upper lip contour 37 Face contour 51 Rn peak 52 Rn + 1 peak 71 Camera 72 Display 73 PC part 74 Database 75 Net line 76 Heart rate measuring device 81 Face Features 82 Skin area

Claims (11)

A camera that measures the person to be measured, a recording unit that records images taken by the camera, a feature point extraction unit that extracts feature points from the recorded image data, and a feature point comparison that compares the extracted feature points. It includes a unit, a data accumulation unit that accumulates the extracted data, a data comparison unit that compares the accumulated data with the data measured this time, and a judgment unit that judges stress from the compared data.
When the person to be measured is in a healthy state, the person to be measured is photographed with the camera before and after the stress is applied, and an image is recorded in the recording unit.
The feature points are extracted by the feature point extraction unit from the photographed image data of the body of the person to be measured, and the feature points are extracted.
The coordinates of these feature points are stored in the standard database and
The data comparison unit represents the difference between the coordinates of the feature points in the healthy state of the person to be measured and the coordinates of the feature points at the time of measurement stored in the standard database as a vector.
The judgment unit is a stress evaluation system that evaluates the stress that the person to be measured receives at the time of measurement from this vector.
There are multiple people to be measured, and the standard database stores the coordinate data of the feature points of the plurality of people to be measured, and it is a healthy person other than the person to be measured to evaluate the stress that the person to be measured is receiving. The vector based on the difference between the coordinates of the feature point in the state and the coordinates of the feature point in the measurement is also obtained.
Identify the vector frequent pattern from the vector obtained from the difference in coordinates.
A stress evaluation system characterized in that a correlation value with the emotion of the person to be measured is obtained by feedback consisting of a questionnaire result from the person to be measured, a medical inquiry result, or a result of the person's declaration. The stress evaluation system according to claim 1, wherein the face of the person to be measured is photographed by the camera, and the feature points of the facial expression of the person to be measured are extracted from the photographed face image data. The stress evaluation system according to claim 1, wherein the whole body of the person to be measured is imaged with the camera, and the characteristic points of the person to be measured are extracted from the captured moving image data. The standard database is composed of a group standard database created from the entire subject, a specific group standard database created from a specific group, and an individual standard database created from the past measurement results of the individual subject. The stress evaluation system according to claim 1, characterized in that The stress evaluation system according to any one of claims 1 to 4, wherein the pulse wave of the person to be measured is measured. A pulse wave measurement unit that measures the pulse wave, a Lorentz plot generation unit that obtains the Lorentz plot of the change in the interval of the peak value of the measured pulse wave, and a stress evaluation unit that evaluates the stress degree from the distribution of the Lorentz plot. The stress assessment system according to claim 5, wherein the stress assessment system includes. The pulse wave measuring unit is a color change measuring unit that measures a change in skin color due to blood flow in the face of the person to be measured measured by the camera, and a pulse wave that calculates a pulse wave from the measured color change. The stress evaluation system according to claim 6, further comprising a calculation unit. The stress evaluation system according to any one of claims 1 to 7, wherein the camera comprises a plurality of cameras. The camera is mounted on the gate to be measured person passes, stress evaluation system according to claim 1, wherein the shooting to be measured person in conjunction with the gate passage of the measured person. The camera is attached to the room where the person to be measured conducts a health examination, a meeting, a meeting, work or work, and is linked to the person to be measured to participate in a health examination, a meeting, participate in a meeting, perform work or perform work. The stress evaluation system according to claim 1, wherein the person to be measured is photographed. The stress evaluation system according to any one of claims 1 to 10, wherein the voice of the person to be measured is recorded by a microphone, feature points are extracted, and the characteristic points are used for stress evaluation.

Images