Biological information detection device and biological information detection method

Abstract

To provide a biological information detection device and a biological information detection method which can detect the biological information at high speed without taking much time for measurement.SOLUTION: A biological information detection device 50 of the present invention includes: a pulse wave extraction unit 20 for extracting pulse waves from facial image information; a real-time reference pulse wave threshold generating unit 32 which obtains a reference pulse wave and a predetermined first threshold being a tolerance range of the reference pulse wave for each pulse wave extracted by the pulse wave extraction unit; a pulse wave determination / selection unit 31 which excludes the pulse wave if the pulse wave extracted by the pulse wave extraction unit is outside of a range of the predetermined first threshold obtained by the rear-time reference pulse wave threshold generation unit; and a biological information calculation unit 30 for calculating biological information from the pulse wave selected by the pulse wave determination / selection unit.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a biological information detection device and a biological information detection method. [Background technology]

[0002] In recent years, the use of wearable devices has been considered as a method for managing health. Some wearable devices are equipped with optical sensors to constantly monitor blood flow and output and transmit various vital information.

[0003] However, managing one's health by wearing a wearable device can be a burden for the worker, and there are cases where the worker cannot work while wearing the wearable device. Due to these issues, it is sometimes desirable to be able to measure health without wearing a wearable device.

[0004] When monitoring blood flow, measurement time is important, and it is desirable to measure within a time that does not bother the user. On the other hand, when focusing on accuracy, a stable measurement system with small error and high accuracy can be constructed by adopting a value obtained by averaging multiple sensor values ​​over time. However, there is a problem in that trying to improve accuracy takes time to measure.

[0005] Patent Document 1 describes a technology for correcting heart rate. In detail, it discloses that "a computer executes the following process: acquiring heart rate data relating to the heart rate state from a first sensor that measures the heart rate of a subject; acquiring exercise data from a second sensor that measures the exercise of the subject; identifying an exercise section within a measurement section from the exercise data; setting a correction section including the exercise section and a predetermined section after the exercise section; identifying a lower heart rate between the heart rate at the start and end of the correction section based on the heart rate data; replacing the heart rate data within the correction section with correction data corrected to the identified heart rate; and detecting a specific behavior of the subject based on the correction data for the correction section and heart rate data for a section within the measurement section other than the correction section." [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6828619 Summary of the Invention [Problem to be solved by the invention]

[0007] According to the above-mentioned prior art, the influence of heart rate due to exercise is corrected, so accuracy is not reduced. However, the correction is performed based on the heart rate at the start and end of the correction target period calculated from the subject's exercise data, and if the subject exercises for a long time, it takes a long time to detect the subject's behavior.

[0008] An object of the present invention is to provide a biological information detection device and a biological information detection method that can detect biological information at high speed without taking a long time for measurement. [Means for solving the problem]

[0009] In order to solve the above problems, the biometric information detection device of the present invention includes a pulse wave extraction unit that extracts a pulse wave from facial image information, a real-time reference pulse wave threshold generation unit that calculates a predetermined first threshold that is an error range between a reference pulse wave and the reference waveform for each pulse wave extracted by the pulse wave extraction unit, a pulse wave judgment and selection unit that excludes the pulse wave extracted by the pulse wave extraction unit if it is outside the range of the predetermined first threshold calculated by the real-time reference pulse wave threshold generation unit, and a biometric information calculation unit that calculates biometric information from the pulse wave selected by the pulse wave judgment and selection unit. [Effects of the Invention]

[0010] According to the biological information detection device and biological information detection method of the present invention, biological information can be detected at high speed. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 10 is a diagram showing a determination flow for separating an error from a pulse wave for determining biological information. [Figure 2] 1A to 1C are diagrams illustrating in detail the biological information detection method according to the embodiment, taking pulse rate as an example. [Figure 3] 1 is a configuration diagram of a biological information detection device according to a first embodiment. [Figure 4] FIG. 10 is a configuration diagram of a biological information detection device according to a second embodiment. [Figure 5] FIG. 10 is a configuration diagram of a biological information detection device according to a third embodiment. [Figure 6] FIG. 10 is a configuration diagram of a biological information detecting device according to a fourth embodiment. [Figure 7] 4 is a flowchart illustrating the operation of the biological information detection device 50. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. First, an overview of biological information detection according to the embodiment will be described. In this embodiment, biometric information detection involves analyzing blood flow in a non-contact manner from a video of the face captured by a camera to extract a pulse wave, and then detecting biometric information such as pulse rate, blood pressure, LF / HF (stress index), and SpO2 (oxygen saturation) based on the pulse wave. To obtain accurate biometric information, it is necessary to separate the pulse wave, which contains errors (due to device-specific factors, measurement environment, and individual differences).

[0013] For this reason, in the biological information detection method of the embodiment, the error and the pulse wave for determining the pulse rate are separated using the judgment flow in Fig. 1. The judgment flow in Fig. 1 judges the pulse wave corresponding to each pulse rate extracted by analyzing the blood flow from the intensity image of reflected light showing the change in hemoglobin concentration detected by near-infrared spectroscopy.

[0014] In step S1, the biological information detection method acquires a pulse candidate to be determined whether or not it is a pulse wave for which a pulse is to be obtained.

[0015] In step S2, the biometric information detection method determines whether the pulse candidate is within the range of pulses a human can take, and if it is within the range of pulses a human can take (Yes in S2), the method proceeds to step S3. If it is not within the range of pulses a human can take (No in S2), the method proceeds to step S5, determines that the pulse candidate is a pulse wave containing error, and ends the process. Here, "a pulse within the range of pulse rates a human can take" refers to the range of pulse wave signal strength (amplitude).

[0016] In step S3, the biological information detection method determines whether the pulse candidate has undergone a sudden change compared to the pulse previous to this candidate, and if there has been no sudden change (No in S3), the method proceeds to step S4. If there has been a sudden change (No in S3), the biological information detection method proceeds to step S5, where the pulse candidate is determined to be a pulse wave containing error, and the process ends. Here, "sudden change" means a large change in pulse rate or pulse wave signal strength.

[0017] In step S4, the biological information detection method determines that the pulse wave of the pulse candidate is not an error but a pulse wave for determining the actual pulse. In other words, if the pulse candidate is within a predetermined range and does not change suddenly, the biological information detection method determines that it is a pulse wave for determining the actual pulse. As will be described in detail later, if it is determined that the pulse wave contains an error, the determined pulse wave (pulse) is excluded from the detection of biological information.

[0018] Next, the biological information detection method of the embodiment will be described in detail with reference to FIG. 2, taking pulse rate as an example. FIG. 2 is a diagram showing a process for detecting a pulse rate by sequentially processing a time-series pulse wave.

[0019] The pulse wave indicated by the dashed line in FIG. 2 is excluded as a pulse wave containing errors based on pulse wave shape determination.

[0020] Next, the interval between the peaks of the pulse wave is calculated as the pulse wave interval Tp_tmp. Then, as shown in the upper graph showing the change over time in pulse wave intervals, pulse wave intervals Tp_tmp (corresponding to the reciprocal of the heart rate) that are equal to or greater than a predetermined value are excluded and set as the valid pulse wave interval Tp_s_ext1. For example, pulse wave intervals Tp_tmp that are equal to or greater than 2 seconds, corresponding to a heart rate of 30 BPM, are excluded (indicated by a dashed circle).

[0021] This corresponds to the determination of whether or not the pulse wave is within the range that a human can obtain in step S2 in Fig. 1. In the above, the waveform indicated by the dashed line is excluded as a pulse wave containing errors by determining the pulse wave shape, but it can also be excluded by determining the pulse wave interval Tp_tmp without determining the pulse wave shape.

[0022] In the biological information detection method of this embodiment, a moving average of the valid pulse interval Tp_s_ext1, excluding pulses that do not occur in normal measurement, is calculated and used as a carefully selected reference Tp_s_ave. As shown in the graph below showing changes in pulse interval over time, a range of ±30% of this reference Tp_s_ave is set as a strict reference range, and values ​​in the valid pulse interval Tp_s_ext1 that exceed this strict reference range are excluded as abnormal values ​​(indicated by dashed circles) and set as the strict pulse interval Tp_s_m.

[0023] This process excludes pulse intervals whose values ​​have changed significantly, which corresponds to determining whether the pulse candidate has changed suddenly in step S3 in Figure 1. The average of the selected strict pulse intervals Tp_s_m is calculated as the reference pulse interval, which is set as the pulse interval reference TpRef. In the biological information detection method according to the embodiment, the reciprocal of the pulse interval reference TpRef calculated as described above is used as the pulse rate.

[0024] In the above, it has been explained that pulse waves determined to be abnormal are excluded. However, it is also possible to assume that a normal pulse wave (a pulse wave without error) exists between pulse wave intervals, and to insert pulse wave information as if there is a pulse wave that falls within the approximate range of values ​​that can be taken for the pulse wave interval length and amplitude of the reference pulse wave, and continue processing the heart rate, etc.

[0025] The configuration of the biological information detection device 50 according to the embodiment will be described in detail below. Example 1

[0026] FIG. 3 is a configuration diagram of a biological information detection device 50 according to the first embodiment. The biological information detection device 50 comprises an image acquisition unit 10, a pulse wave extraction unit 20, a biological information detection unit 51, and an output unit 40. The biological information detection device 50 is a device that analyzes an image of a human face, extracts a pulse wave, and detects biological information of the human.

[0027] The image acquisition unit 10 is a unit for acquiring image information of a human face, and may use a color camera or a web camera that can be connected to a personal computer. The image acquisition unit 10 may also use a method for reading content containing image information other than a camera. For example, the image acquisition unit 10 may read a video file and read a signal equivalent to the video information captured by a camera. Instead of a video file, the image acquisition unit 10 may read a signal extracted from the video and saved.

[0028] The pulse wave extraction unit 20 extracts information corresponding to the blood flow pulse wave from the image information of the human face acquired by the image acquisition unit 10. Blood flows through subcutaneous capillaries, arteries, and veins, and changes in the color of the blood flow are visible through the skin surface. The pulse wave extraction unit 20 acquires these changes and extracts the blood flow pulse wave.

[0029] Specifically, the greater the blood flow, the greater the amount of red reflected light and the less green reflected light there is. Therefore, the pulse wave extraction unit 20 separates the color information from the video image information and detects increases or decreases in red and green. The pulse wave extraction unit 20 then extracts the pulse wave from this information.

[0030] The biological information detection unit 51 is composed of a real-time reference pulse wave threshold generation unit 32, a pulse wave judgment / selection unit 31, and a biological information calculation unit 30, and calculates biological information from the pulse wave information extracted by the pulse wave extraction unit 20 and notifies the output unit 40 of the biological information.

[0031] The real-time reference pulse wave threshold generator 32 determines the pulse wave interval and the peaks and valleys of each pulse wave, and calculates a reference pulse wave and a predetermined threshold (first threshold) that is an error range for the reference waveform. The pulse wave shape is checked in order from the first waveform, and after confirming that a sufficient amplitude and pulse wave interval for the living body have been obtained, data is collected as the reference pulse wave. The pulse wave interval and amplitude are calculated from this first collected pulse wave data, and a biologically acceptable error range is set as the threshold (first threshold). This threshold is the pulse wave interval and pulse wave amplitude, and the pulse wave amplitude is set as the difference between the peak value of the peak and the peak value of the valley.

[0032] As explained in FIG. 2, if the pulse wave input in real time is within a predetermined threshold value (first threshold value) for each threshold value output from the real-time reference pulse wave threshold value generating unit 32, the pulse wave judgment / selection unit 31 stores the data as a normal pulse wave (a pulse wave containing no error), and if the pulse wave input in real time is outside the predetermined threshold value (first threshold value) range, the pulse wave is judged to be an abnormal pulse wave containing an error and excludes the data.

[0033] More specifically, the pulse wave judgment / selection unit 31 determines the valid pulse wave interval Tp_s_ext1 explained in Fig. 2 and performs processing corresponding to step S2 in Fig. 1, as well as processing corresponding to step S3 in Fig. 1 below. That is, as explained in Fig. 2, the pulse wave judgment / selection unit 31 excludes pulse waves whose valid pulse wave interval Tp_s_ext1 falls outside the strict reference range (a range of ±30% of the reference Tp_s_ave) generated by the real-time reference pulse wave threshold generation unit 32 as the threshold for the pulse wave interval of the reference pulse wave.

[0034] The pulse wave determination / selection unit 31 analyzes the shape of the pulse wave extracted by the pulse wave extraction unit 20, extracts feature points of the shape, and determines the type of pulse wave. The feature points of the pulse wave shape are the signal value and time of the peak of the waveform's peak and valley peak, and the amount and rate of change in the rise before the peak and the fall after the peak. Feature points are extracted not only for one pulse wave shape, but also for multiple consecutive pulse wave shapes.

[0035] Examples of pulse wave types include a pulse wave that is symmetrical with respect to the peak, a pulse wave whose peak is biased to one side, a pulse wave with a low peak, a pulse wave with a high peak, a pulse wave with an undulating peak, etc. The pulse wave determination / selection unit 31 may determine whether the pulse wave is within or outside a threshold value (first threshold value) of a reference pulse wave based on the type of pulse wave, determine the attribute, and determine whether or not to acquire it as a pulse wave.

[0036] As described above, pulse wave judgment / selection unit 31 has the function of selecting necessary pulse waves from unnecessary pulse waves and switching between processes to omit unnecessary processing. The usefulness of pulse wave judgment / selection unit 31 will be explained in more detail below.

[0037] If the pulse wave is continuously within the reference pulse wave threshold (first threshold), it is likely to have a well-balanced pulse wave shape, and the pulse wave signal is likely to be free of errors. This pulse wave information is considered to be highly reliable. It is desirable to use such pulse waves for measurement and to continue updating the reference pulse wave.

[0038] On the other hand, if a pulse wave is judged to have a low peak, and if the peak is extremely low, it is considered that the pulse wave was not measured correctly. If such pulse wave information is used, it will cause errors in the detection of biological information.

[0039] Even if the pulse wave contains errors and cannot be measured correctly, a person's pulse wave is a continuous pulse, and since this is detected, pulse data can be obtained continuously and repeatedly. In other words, when trying to measure a pulse, even if there are errors in the pulse wave and the pulse wave intervals vary, by measuring and averaging a large amount of data, the pulse for that averaged time period can be measured.

[0040] Measuring and averaging a large amount of data generally acts as a low-pass filter, eliminating sudden noise. However, this type of error reduction process requires a large amount of data to achieve sufficient accuracy, which results in long data collection and processing times.

[0041] The pulse wave judgment and selection unit 31 can obtain highly accurate results in a short time by using only pulse waves that are likely to be within the range of the reference pulse wave threshold (first threshold) from among the extracted pulse waves. If valid data arrives consecutively, time smoothing may be performed a smaller number of times.

[0042] Furthermore, without the real-time reference pulse wave threshold generator 32, pulse wave data acquisition must wait until multiple normal pulse wave data arrives, which takes time to acquire the data before determining the threshold. Alternatively, to obtain the desired accuracy even for pulse wave intervals, noise removal processing such as time smoothing must be performed over a longer period of time. Furthermore, using pulse wave information containing noise introduces errors into the calculation results of biological information, so noise removal also takes time.

[0043] By providing the real-time reference pulse wave threshold generating unit 32 and the pulse wave determining / selecting unit 31 of the first embodiment, it is possible to prevent the measurement time from becoming too long.

[0044] The biological information calculation unit 30 calculates the heart rate, blood pressure, and the stress index LF / HF as biological information based on the pulse wave from which abnormal pulse waves containing errors have been removed using the real-time reference pulse wave threshold generation unit 32 and the pulse wave judgment / selection unit 31.

[0045] Heart rate can be calculated by measuring the time interval between pulse waves and calculating how many times the pulse beats in 60 seconds. The mean blood pressure is calculated by averaging the peaks and valleys of the pulse wave over a specified period. If the mean blood pressure is measured simultaneously with a blood pressure monitor, a medical measuring device, and the blood pressure calculation method described above is set based on this value, it becomes possible to measure blood pressure without contact. The stress index LF / HF is calculated by frequency analysis of fluctuations in pulse wave intervals.

[0046] The output unit 40 receives the biometric information calculated by the biometric information calculation unit 30 and presents it in a form that can be seen by a person using a display, a printer, etc. Alternatively, the biometric information may be stored in a database server instead of being presented directly to a person.

[0047] The functions of the biological information detection unit 51 and the pulse wave extraction unit 20 in the first embodiment may be realized by a server-side program, and a biological information measurement system may be configured in which the image acquisition unit 10 and the output unit 40 are connected via a network. This allows an image of a person to be acquired and the biological information thereof to be measured simply and without contact. Example 2

[0048] Next, the configuration of a biological information detection device 50 according to a second embodiment will be described with reference to FIG. 3 in that the biological information detection unit 51 includes a real-time reference pulse wave threshold generator 32 and a biological pulse wave setting unit 33. The other components are the same as those in FIG. 3, and therefore will not be described here.

[0049] The biological pulse wave setting unit 33 sets an initial value of a threshold used by the real-time reference pulse wave threshold generating unit 32 to determine whether the pulse wave extracted by the pulse wave extracting unit 20 is within a normal range of pulse wave amplitude and pulse wave interval as a pulse wave for detecting human biological information, and whether the pulse wave amplitude and pulse wave interval are within a range that can be taken in an abnormal state.

[0050] As a result, even if the initially input pulse wave is abnormal, since a threshold value is set to an initial value, the input pulse wave is instantly removed by the pulse wave judgment / selection unit 31. When a normal pulse wave (a pulse wave containing no error) is input, pulse wave data is instantly collected by the pulse wave judgment / selection unit 31, and biological information is calculated by the biological information calculation unit 30. Example 3

[0051] Next, the configuration of a biological information detection device 50 according to a third embodiment will be described with reference to FIG. The biometric information detection device 50 in Fig. 5 differs from the biometric information detection device 50 in Fig. 3 and Fig. 4 in that the biometric information detection unit 51 includes a face detection unit 34 having a real-time face contour extraction unit 35. The other configurations are the same as those in Fig. 3 and Fig. 4, and therefore will not be described here.

[0052] The real-time face contour extraction unit 35 extracts a face contour from the image information of the human face acquired by the image acquisition unit 10 using various information such as contrast ratio, color difference, motion vector, color information, and focal depth.

[0053] The face detection unit 34 determines whether or not face detection is measurable and identifies facial parts using the face contour extracted by the real-time face contour extraction unit 35, and also determines the amount of face movement, such as determining whether the amount of movement in all directions of the face contour (hereinafter referred to as face movement amount) is within a predetermined threshold range or not. As a result, the biometric information detection device 50 of the third embodiment determines whether the face is moving or vibrating slightly, detects face movement that is largely caused by pulse wave errors, excludes the extracted pulse wave, and calculates biometric information.

[0054] Specifically, the face detection unit 34 calculates the amount of blur (amount of movement) in all directions of the face contour extracted by the real-time face contour extraction unit 35, and sets the largest value as the amount of face movement. The predetermined threshold (second threshold) range determined by the face detection unit 34 is the amount of face movement that can be considered not to affect the detection of changes in color information of the facial image information that changes in response to blood flow velocity. In other words, the predetermined threshold (second threshold) range is preferably calculated from image information of a stationary face for pulse wave detection, but corresponds to the allowable range of pulse wave error caused by facial movement.

[0055] Furthermore, the face detection unit 34 sets a judgment threshold (second threshold) for the amount of facial movement at which an abnormality in the pulse wave occurs, calculates the amount of facial movement from the face contour extracted from the video information, and judges whether the calculated amount of facial movement exceeds the judgment threshold (second threshold). If the calculated amount of facial movement exceeds the judgment threshold, the face detection unit 34 immediately issues an alarm to warn the user. This makes it possible to immediately fix the face and start or resume measurement, which leads to a reduction in the time required to detect biological information and improves real-time performance.

[0056] Furthermore, the biological information detection device 50 utilizes the information classified by the pulse wave judgment / selection unit 31 to determine the stability of measurement by linking the number and detection order of pulse waves that are likely to be measured correctly and those that are likely to have a large error with the amount of face movement, and then determining the result. For example, if the amount of face movement is within the face movement amount threshold (second threshold) and X consecutive pulse waves are within the reference pulse wave threshold (first threshold), the pulse wave is determined to be stable, and biological information detection continues. Conversely, if a certain amount of face movement is outside the face movement amount threshold (second threshold), the pulse wave is determined to have a large error and cannot be used as is.

[0057] In this way, by determining the stability of the pulse wave, it is possible to determine whether to continue detecting biological information or to urge the user to stop detection and restart. This function can help the user to quickly achieve accurate detection of biological information. Example 4

[0058] Next, the configuration of a biological information detection device 50 according to a fourth embodiment will be described with reference to FIG. The biological information detection device 50 in Fig. 6 differs from the biological information detection devices 50 in Fig. 3, Fig. 4, and Fig. 5 in that the biological information detection unit 51 includes a weighting determination unit 36. The other configurations are the same as those in Fig. 3, Fig. 4, and Fig. 5, and therefore will not be described here.

[0059] The weighting determination unit 36 ​​weights the determination information of whether the amount of movement of the face contour of the face detection unit 34 is within or outside the threshold (second threshold) of the amount of movement of the face and the determination information of whether the pulse wave is within or outside the threshold (first threshold) of the reference pulse wave generated by the real-time reference pulse wave threshold generation unit 32 using the weighting determination criterion setting unit 37, and determines whether to remove the pulse wave.

[0060] The weighting determination unit 36 ​​weights each of the two pieces of determination information to make a more accurate determination and improve the accuracy of calculating the biometric information. For example, even if the amount of face movement in the face detection unit 34 is within a threshold (second threshold), if the pulse wave is NG according to the determination criterion of being inside or outside the threshold (first threshold) of the reference pulse wave, it is better to remove this pulse wave, so the weighting of the latter determination information is increased. Conversely, even if the amount of face movement is outside the threshold (second threshold), if the pulse wave is OK (within the threshold) according to the determination criterion of being inside or outside the threshold (first threshold) of the reference pulse wave, the weighting of the latter determination information is increased to continue acquiring a normal pulse wave (a pulse wave without error).

[0061] As described above, by providing the weighting determination unit 36, it is possible to weight the amount of face movement and the pulse wave so that priority is given to the pulse wave being within the threshold (first threshold) when determining the threshold value. Therefore, when the pulse rate is stable without fluctuations over a long period of time, it is possible to calculate the biometric information more stably.

[0062] With the above configuration, the biological information detecting device 50 of the fourth embodiment can achieve real-time biological information detection while ensuring accuracy in a non-contact manner. Furthermore, if the stability of detection itself is impaired, the biological information detecting device 50 warns the user to urge stable detection, thereby providing high usability.

[0063] Next, the operation of the biological information detection device 50 of the embodiment will be described with reference to the flowchart of FIG. The flow chart of FIG. 7 shows the operation of the biological information detection device 50 shown in FIG. 5 when extracting one pulse wave.

[0064] In step S71, the image acquisition unit 10 acquires image information of a human face captured by a color camera.

[0065] In step S72, the face detection unit 34 calculates the amount of movement in all directions of the face contour extracted by the real-time face contour extraction unit 35, sets the maximum value as the amount of face movement, and performs a face movement determination to see if the amount of face movement is within a predetermined threshold (second threshold). If the amount of face movement is within the predetermined threshold (OK in S72), the process proceeds to step S73. If the amount of face movement is not within the predetermined threshold (NG in S72), the process returns to step S71, and is repeated until the amount of face movement (amount of movement in all directions of the face contour) becomes small. At this time, an alert may be output to the user to stay still.

[0066] In step S73, the pulse wave extracting section 20 extracts a pulse wave from the image information of the human face acquired by the image acquiring section 10.

[0067] In step S74, the real-time reference pulse wave threshold generating unit 32 generates, in real time, a threshold (first threshold) of the interval of the reference pulse wave for determining whether or not to remove the pulse wave extracted in step S73, from the reference pulse wave obtained from the previously extracted pulse wave. The threshold at the start of biological information detection is set by the biological pulse wave setting unit 33.

[0068] In step S75, the real-time reference pulse wave threshold generating unit 32 generates, in real time, thresholds for the peak and trough values ​​of the reference pulse wave to determine whether or not to remove the pulse wave extracted in step S73. The threshold at the start of biological information detection is set by the biological pulse wave setting unit 33.

[0069] In step S76, pulse wave judgment / selection unit 31 judges in real time whether the pulse wave extracted in step S73 satisfies the threshold value (first threshold value) generated in steps S74 and S75. If the threshold value is satisfied (OK in S76), pulse wave judgment / selection unit 31 proceeds to step S78, and if the threshold value is not satisfied (NG in S76), it proceeds to step S77.

[0070] In step S77, the pulse wave determining / selecting unit 31 removes the pulse wave extracted in step S73 as an abnormal pulse wave containing an error.

[0071] In step S78, the biological information detection unit 51 updates the reference pulse wave with a pulse wave that does not include an error.

[0072] In step S79, the biological information calculation unit 30 calculates biological information based on the pulse wave from which the abnormal pulse wave containing errors has been removed.

[0073] The biological information detection device 50 of the embodiment detects biological information for each pulse wave extracted from video information according to the flow chart of FIG. 7, so that measurement does not take much time and biological information can be detected quickly.

[0074] Furthermore, the present invention is not limited to the above-described examples, and various modifications are included. The above-described examples have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. [Explanation of symbols]

[0075] 10 Video acquisition unit 20 Pulse wave extraction unit 30 Biometric information calculation unit 31 Pulse wave judgment and selection unit 32 Real-time reference pulse wave threshold generator 33 Biopulse wave setting unit 34 Face detection unit 35 Real-time face contour extraction unit 36 Weighting judgment unit 37 Weighting criteria setting section 40 Output section 50 Biometric information detection device 51 Biometric information detection unit

Claims

1. a pulse wave extraction unit that extracts a pulse wave from facial image information; a real-time reference pulse wave threshold generating unit that calculates a reference pulse wave and a predetermined first threshold that is an error range for the reference waveform for each pulse wave extracted by the pulse wave extracting unit; a pulse wave determination and selection unit that excludes the pulse wave extracted by the pulse wave extraction unit if the pulse wave is outside the range of the predetermined first threshold calculated by the real-time reference pulse wave threshold generation unit; a biological information calculation unit that calculates biological information from the pulse wave selected by the pulse wave determination / selection unit; A biological information detection device comprising:

2. 2. The biological information detection device according to claim 1, the real-time reference pulse wave threshold generating unit calculates a pulse wave interval and a pulse wave amplitude from the pulse wave interval and the peaks and valleys of the pulse wave, and sets the first threshold value within a biologically acceptable error range. Biometric information detection device.

3. 3. The biological information detection device according to claim 1, further comprising: a face detection unit that calculates the amount of movement of the face contour in all directions from the face contour extracted from the video information, sets the maximum value as the amount of face movement, and determines that the pulse wave error is large if the amount of face movement exceeds a predetermined second threshold; A biological information detection device comprising:

4. 4. The biological information detecting device according to claim 3, a weighting determination unit that weights information on whether the amount of face movement detected by the face detection unit is within the second threshold and information on whether the pulse wave detected by the pulse wave determination and selection unit is within the first threshold of a reference pulse wave and determines whether to remove the pulse wave; A biological information detection device comprising:

5. extracting a pulse wave from facial image information; determining a reference pulse wave and a predetermined first threshold value, which is an error range for the reference waveform, for each of the extracted pulse waves; excluding the extracted pulse wave if it is outside the range of the predetermined first threshold value of the reference pulse wave; calculating biological information from the pulse wave; A biological information detection method comprising:

Images