Biological detection system, biological detection method, and program

Abstract

This invention is provided with: an averaging processing unit for acquiring video image data for a face image obtained by capturing the face of a subject, and for a specific region of the acquired video image data, averaging, across pixels, the difference components for primary color components in the specific region every given period; a variation acquisition unit for acquiring variations in each primary color component averaged by the averaging processing unit; and a detection unit for detecting, on the basis of the variation amounts acquired by the variation acquisition unit, blood flow or activity based on blood flow. It is thereby possible to detect blood flow or activity such as autonomic nerve activity in an easy, highly accurate, and contactless manner without special equipment.

Description

Living body detection system, living body detection method, and program

[0001] The present invention relates to a biological detection system, a biological detection method, and a program for detecting biological conditions such as autonomic nervous activity of a subject in a non-contact manner.

[0002] In recent years, there has been an increasing need for emotion estimation in the workplace to improve people's lives, such as by measuring the fatigue and concentration levels of workers and incorporating appropriate rest to improve the workplace environment. To measure fatigue and concentration levels, various emotion estimation methods have been developed using biometric information such as heart rate, pulse wave, electroencephalogram, eye movement, and facial thermal imaging. In particular, the facial thermal imaging method, unlike other methods, can be estimated using a non-contact far-infrared camera. This method is considered to be a useful technique because it can be used without stressing the subject or restricting their behavior.

[0003] Patent Document 1 describes a driver monitoring system that uses thermal images or the like to obtain the driver's emotions.

[0004] Japanese Patent Application Laid-Open No. 2021-146214

[0005] Conventional emotion estimation using facial thermal images estimates autonomic nervous activity by taking the temperature difference between the nose, which is an indicator of the sympathetic nervous system in autonomic nervous activity, and the forehead, which is less affected by autonomic nervous activity. Skin temperature is the antagonistic temperature of the conductive heat of deep body temperature, changes in blood flow under the skin, and the environmental temperature. Therefore, by taking the temperature difference between the nose and forehead, it is possible to obtain only changes in peripheral blood flow under the skin of the nose, excluding changes in external temperature and deep body temperature.

[0006] When an unpleasant stimulus is received, the temperature of the nose drops and the temperature of the mouth rises. However, even when an unpleasant stimulus is received, the average temperature of the entire face does not change significantly. Therefore, it is known that it is possible to estimate "comfort or discomfort" from changes in peripheral blood flow under the skin of the nose, using the variance value of the facial thermal image as an index based on changes in temperature distribution due to emotional changes. However, such conventional methods require a far-infrared camera to capture thermal images, making it difficult to easily detect autonomic nervous activity using existing systems, and there are problems with the cost required to prepare a far-infrared camera.

[0007] Furthermore, we have mentioned here the problem of detecting autonomic nervous activity, but it has been difficult to accurately detect blood flow from facial images in the first place.

[0008] The object of the present invention is to provide an autonomic nervous activity detection system, an autonomic nervous activity detection method, and a program that can easily detect biological conditions such as autonomic nervous activity in a non-contact and highly accurate manner without the need for special equipment.

[0009] The biometric detection system of the present invention includes an averaging processing unit that acquires video data of a facial image of a subject's face and averages, for a specific region of the acquired video data, all pixels of primary color components within the specific region at regular intervals; a fluctuation acquisition unit that obtains the fluctuation for each primary color component averaged by the averaging processing unit; and a detection unit that detects blood flow or activity based on that blood flow based on the amount of fluctuation obtained by the fluctuation acquisition unit.

[0010] The present invention also provides a living body detection method in which an information processing device performs arithmetic processing based on video data of a facial image of a subject's face to detect blood flow or activity based on that blood flow. The arithmetic processing performed by the information processing device includes an averaging process for averaging, over all pixels, difference components of primary color components within a specific region of the video data of the subject's face at regular intervals, a fluctuation acquisition process for obtaining fluctuations in the difference components of the primary color components averaged by the averaging process, and a detection process for detecting blood flow or activity based on that blood flow based on the amount of fluctuation obtained by the fluctuation acquisition process. The present invention also provides a program for causing a computer to execute each step of the above-described living body detection method.

[0011] According to the present invention, a process can be performed to accurately detect blood flow in blood vessels at specific locations or activity based on that blood flow from moving images of a face.

[0012] 1 is a block diagram showing an example of a living body detection system according to an embodiment of the present invention. FIG. 2 is a block diagram showing an example of a hardware configuration when the living body detection system according to an embodiment of the present invention is configured with a computer. FIG. 3 is a flowchart showing the flow of living body detection processing according to an embodiment of the present invention. FIG. 4 is a diagram explaining estimation of autonomic nerve activity as viewed from a subcutaneous structure according to an embodiment of the present invention. FIG. 5 is a diagram showing an example of a detection state (an example comparing RB components and R+G+B components) according to an embodiment of the present invention. FIG. 6 is a diagram showing an example of a section extracted from the detection state shown in FIG. 5 where the influence of changes in light intensity is small. FIG. 7 is a diagram showing an unrestricted image (left) near the nose and an image (right) where the section extracted in the example shown in FIG. 6 where the influence of changes in light intensity is small. FIG. 8 is a diagram showing examples of pixel values ​​of the unrestricted RB component and the restricted RB component when shaking the head. FIG. 9 is a diagram showing the RB+G+B component and the RB component detected under certain conditions. FIG. 10 is a diagram showing the RB+G+B component and the RB component detected under the same conditions as FIG. 10. FIG. 11 is a diagram showing the RB+G+B component and the RB component detected under the same conditions as FIG. 10. FIG. 12 is a diagram showing an example of landmark detection from a face image according to an embodiment of the present invention. 1A and 1B are diagrams showing examples (example 1 and example 2) of analysis results according to an embodiment of the present invention;

[0013] An embodiment of the present invention (hereinafter referred to as "this embodiment") will be described below with reference to the accompanying drawings. [System Configuration] Fig. 1 is a functional block diagram showing the configuration of the processing performed by an autonomic nerve activity detection system 100 of this embodiment. The autonomic nerve activity detection system 100 of this embodiment captures image data captured by a camera 1. The image data is an image of the face of a subject whose pulse wave is to be detected. The image data captured by the camera 1 is moving image data at a constant frame rate, and the image data for each frame is composed of pixel data of the primary colors red (R), green (G), and blue (B). The frame rate of the image data is, for example, 30 frames per second, and each pixel in each frame represents the brightness value of red (R), green (G), or blue (B) at a predetermined gradation.

[0014] Image data captured by the autonomic nerve activity detection system 100 is supplied to an RGB component acquisition unit 101, which acquires the red, green, and blue components individually. The red, green, and blue components acquired by the RGB component acquisition unit 101 are each supplied to an averaging processing unit 102.

[0015] The averaging processor 102 performs averaging processing on a specific region of the acquired video data at regular intervals, averaging all pixels of the primary color components within the specific region. Here, the specific region of the video data is the nose region. Note that the nose region is an example; any other region with thin skin and subcutaneous fat where changes in blood flow can be easily observed may be used. Furthermore, the (R-B) component is obtained by taking the difference between the red (R) component and the blue (B) component, and the (R-B) component of the nose region is averaged (smoothed) for all pixels within the region at regular intervals. The averaging period is preferably selected between approximately 5 seconds and 1 minute. For example, the averaging processor 102 averages every 20 seconds or 1 minute. Furthermore, when averaging the (R-B) component, the averaging processor 102 extracts only stable regions with little change in light intensity and averages only those stable regions with little change in light intensity. The process of extracting only stable areas with little change in the amount of light will be described later with reference to FIGS.

[0016] The averaging result obtained by the averaging processing unit 102 is then supplied to the fluctuation acquiring unit 103, which performs fluctuation acquiring processing on the averaging result. The fluctuation acquiring unit 103 acquires fluctuations in the average value for each fixed time period. Specifically, it acquires fluctuations such as a state in which the average value for each fixed time period remains almost unchanged, a state in which the average value increases, a state in which the average value decreases, and the like.

[0017] The information on the fluctuation status acquired by the fluctuation acquisition unit 103 is supplied to the detection unit 104. The detection unit 104 performs a process of detecting the stress state of the subject based on the information on the fluctuation status. The detection result (evaluation result) of the stress state detected by the detection unit 104 is output from the output unit 105.

[0018] [Example of Hardware Configuration of Autonomic Nervous Activity Detection System] Fig. 2 shows an example of the hardware configuration of an autonomic nervous activity detection system 100 configured with a computer, which is an information processing device. The computer (autonomic nervous activity detection system 100) shown in Fig. 2 includes a CPU (Central Processing Unit) 100a, a main memory unit 100b, a non-volatile storage 100c, a network interface 100d, an image input unit 100e, an output unit 100f, and an operation unit 100g, which are all connected to a bus.

[0019] The CPU 100a is an arithmetic processing unit that reads and executes program code of software that realizes the functions performed by the autonomic nerve activity detection system 100 from the main memory unit 100b or the nonvolatile storage 100c. The CPU 100a reads the program code from the main memory unit 100b or the nonvolatile storage 100c and executes arithmetic processing in a work area of ​​the main memory unit 100b, thereby configuring various processing function units in the main memory unit 100b. For example, the main memory unit 100b is configured with an RGB component acquisition unit 101, an averaging processing unit 102, a fluctuation acquisition unit 103, and a detection unit 104 shown in FIG. 1 .

[0020] The nonvolatile storage 100c may be a large-capacity information storage medium such as a hard disk drive (HDD), a solid state drive (SSD), a memory card, etc. The nonvolatile storage 100c stores software that realizes the functions of the autonomic nerve activity detection system 100 and data obtained by executing the software.

[0021] The network interface 100d uses, for example, a network interface card (NIC) and transmits and receives data to and from other devices. The image input unit 100e performs input processing (image acquisition processing) of image data from the camera 1. The output unit 100f is connected to a display 2, and data such as images showing detection results of autonomic nervous activity, such as stress evaluation results, are output from the output unit 100f. These detection results and evaluation results are then displayed on the display 2. The detection results and evaluation results may be transmitted to an external device via the network interface 100d. The operation unit 100g accepts operations from devices such as a keyboard and mouse operated by the operator of this computer.

[0022] [Flow of Autonomic Nervous Activity Detection Processing] Figure 3 is a flowchart showing the flow of autonomic nervous activity detection processing performed by the autonomic nervous activity detection system 100 of this example with the configuration shown in Figure 1. First, the image input unit 100e (Figure 2) of the autonomic nervous activity detection system 100 captures facial image data of the subject photographed by the camera 1 (step S11). Then, the RGB component acquisition unit 101 separately acquires the same image data for the primary colors of red (R), green (G), and blue (B) (step S12).

[0023] The primary color components of the image data obtained by the RGB component acquisition unit 101 are supplied to the averaging processing unit 102, which performs averaging processing for the specific region, i.e., the nose region, at regular intervals to average the primary color components across all pixels (step S13). That is, averaging processing is performed for the nose region, which is a specific region of the moving image. Here, the (R-B) component is obtained as the color component by taking the difference between the red (R) component and the blue (B) component, and this (R-B) component is averaged across all pixels within the region at regular intervals (e.g., every 20 seconds, every minute, etc.).

[0024] Then, based on the results of averaging performed by the averaging processor 102, the fluctuation acquisition unit 103 performs a fluctuation acquisition process for the averaged results at regular intervals (step S14). In step S14, the fluctuation acquisition unit 103 acquires fluctuation conditions, such as a state in which the average value for each regular interval remains almost unchanged, a state in which the average value increases, or a state in which the average value decreases. Here, the (R-B) component is extracted (step S14a), and only stable portions that are little affected by changes in light intensity are extracted (step S14b), and the fluctuation conditions are acquired from the resulting average values. Note that the extraction of the (R-B) component and the extraction of only stable portions may be performed by the averaging process.

[0025] Furthermore, the detection unit 104 performs a stress detection process for the subject (step S15) based on the information on the fluctuation status acquired by the fluctuation acquisition unit 103. The stress detection result (evaluation result) obtained by the detection unit 104 is output from the output unit 105.

[0026] [Explanation of Each Process for Detecting Autonomic Nerve Activity] Figure 4 is a diagram illustrating the difference in penetration depth of each primary color wavelength into human skin. As shown in Figure 4, the red (R) component, green (G) component, and blue (B) component have the highest penetration depth into the skin, respectively. That is, for the blue component, light reflected from the epidermis of the skin is captured in the image captured by camera 1. For the green component, light reflected from the dermis, which is deeper than the epidermis, is captured in the image captured by camera 1. Capillaries are located at the depth of this dermis. Furthermore, for the red component, light reflected from the subcutaneous tissue, which is deeper than the dermis, is captured in the image captured by camera 1. Arteries and veins are located in the subcutaneous tissue.

[0027] Here, in the averaging processing unit 102 of this example, the (R-B) component is obtained by taking the difference between the red (R) component and the blue (B) component, and the (R-B) component is considered to contain information about the entire blood vessel excluding the luminance component. Note that if the (G-B) component is obtained by taking the difference between the green (G) component and the blue (B) component, the (R-B) component is considered to contain information about the entire blood vessel excluding the luminance component. In the following description, an example will be described in which the (R-B) component is obtained to obtain information about the entire blood vessel. However, it is also possible to obtain information about the entire blood vessel by obtaining the (G-B) component to perform stress evaluation, etc. However, as will be described below, it is preferable to obtain information about the entire blood vessel by obtaining the (R-B) component.

[0028] Therefore, it is considered that the averaging processing unit 102 can obtain information on the blood flow rate of the entire blood vessel by obtaining the (R-B) component. In this example, the autonomic nervous activity detection system 100 obtains information on the blood flow rate of the entire blood vessel in this way, and estimates (detects) the autonomic nervous activity of the subject (examinee) from changes in this blood flow rate.

[0029] To explain in more detail how blood flow in blood vessels is obtained, it is believed that light that penetrates deep into the skin is absorbed by the blood through the blood vessels, and the unabsorbed light appears on the surface of the face as reflected light. Therefore, assuming that the value of each color component obtained from the captured image represents the amount of reflected light, it is necessary to know how much is absorbed in the blood. It is believed that the component that absorbs light in blood is hemoglobin. Based on this idea, the characteristics of blood that absorbs more light can be identified from the absorptivity of hemoglobin for each color of light.

[0030] For example, in the red light wavelength range of 610nm to 780nm, deoxygenated hemoglobin (Hb) has a higher absorption rate than oxygenated hemoglobin (HbO2). This means that when there is a lot of blood flowing from arteries, the incident red light is not absorbed as much, and the amount of reflected light increases. On the other hand, when there is a lot of blood flowing from veins, more is absorbed, and the reflected light decreases.

[0031] Furthermore, in the range of 500 nm to 570 nm, which is considered the wavelength of green light, both oxygenated hemoglobin and deoxygenated hemoglobin have similar light absorption rates. It is believed that the measurement of capillaries is not the ratio of arterial to venous blood flow, but rather the total blood flow through the capillaries. It is believed that an increase in capillary blood flow increases the amount of absorption of incident green light and decreases the amount of reflection. Conversely, a decrease in capillary blood flow decreases the amount of absorption and increases the amount of reflection. Using these principles, in this example, the (R-B) component is acquired to detect the blood flow throughout the blood vessels, allowing for estimation (detection) of autonomic nervous activity.

[0032] Next, we will explain the process performed by the averaging processor 102 to extract areas of the (R-B) component that are less affected by changes in light intensity. FIG. 5 is a diagram showing the distribution of luminance values ​​of each pixel in one frame of an image of the nose. The horizontal axis in FIG. 5 represents the (R+G+B) component, with the left end of the horizontal axis corresponding to black and the right end corresponding to white. The vertical axis in FIG. 5 represents the (R-B) component. In this case, the luminance values ​​of each color R, G, and B range from 0 to 255, and the (R+G+B) component ranges from 0 to 765.

[0033] 5, in the range where the (R+G+B) component is close to 0, the distribution of the (R-B) component has a slope indicated by arrow Da. Similarly, in the range where the (R+G+B) component is close to the maximum value (765), the distribution of the (R-B) component has a slope indicated by arrow Db. This range of slope is subject to significant fluctuations due to changes in the amount of light in the image, and is therefore not suitable for the processing of this example.

[0034] Therefore, in this example, as shown in Figure 6, an intermediate range Dx, which is less affected by changes in light intensity, is extracted for the (R-B) component, excluding a specific range near the minimum value of the (R+G+B) component and a specific range near the maximum value. For example, when the (R+G+B) component ranges from 0 to 765, the (R+G+B) component is limited to a range from 115 to 600. However, this value is just an example, and it is preferable to impose appropriate restrictions on the range near the minimum value and the range near the maximum value of the (R+G+B) component.

[0035] Figure 7 shows an example of an image of a nose that has been restricted by the (R+G+B) component values. The left side of Figure 7 is an example of an image without any restrictions. The right side of Figure 7 is an example in which the (R+G+B) components have been restricted from 115 to 600, with the restricted areas appearing black. As can be seen from Figure 7, by restricting the values, bright areas and shadow areas due to the influence of external light, etc. are excluded, and only pixel values ​​suitable for analysis in the processing of this example are extracted.

[0036] Figure 8 shows that the limitations of this example work effectively. The horizontal axis of Figure 8 is time (seconds), and the vertical axis is the value of the (RB) component. Figure 8 shows an example of how the (RB) component changes in the nose area when the subject shakes their head from side to side.

[0037] In FIG. 8, (R-B) ａｌｌ The characteristic shown is the (RB) component obtained from all the (R+G+B) components. ｌｉｍｉｔ The characteristic shown as follows is the (R-B) component obtained from the range of 115 to 600 of the (R+G+B) components. As shown in FIG. 8, the characteristic (R-B) ｌｉｍｉｔ The fluctuation is relatively small even when the subject shakes their head from side to side and the reflection state of the external light changes. ａｌｌ In the case of (2), when the subject shakes their head from side to side, the (RB) component changes relatively significantly due to the large influence of external light. Therefore, by limiting the brightness range as in this example, stable analysis that is not influenced by external light or shadows becomes possible.

[0038] [Example of Why Using the (R-B) Component is Most Desirable] In this example, the analysis is performed using the (R-B) component as the color difference component. The reason why using the (R-B) component is most desirable will be explained with reference to FIGS. 9 to 11.

[0039] 9 to 11 all show the distribution of the same nose image, with the horizontal axis in each of these figures representing the (R+G+B) component. The vertical axis in FIG. 9 represents the (R-B) component, the vertical axis in FIG. 10 represents the (R-G) component, and the vertical axis in FIG. 11 represents the (G-B) component. Comparing these figures, it can be seen that the (R-B) component in FIG. 9 falls within a substantially constant range, making it a signal component suitable for analysis. On the other hand, for the (R-G) component in FIG. 10, the pixel values ​​are spread over a wider range than in the example in FIG. 9, indicating that the analysis results are somewhat less accurate than those for the component in FIG. 9. Furthermore, for the (G-B) component in FIG. 11, the pixel values ​​are spread over an even wider range than in the examples in FIGS. 9 and 10, indicating that the analysis results are even less accurate than in the other cases.

[0040] FIG. 12 is a diagram illustrating the process by which the averaging processor 102 extracts a specific region, the nose region, from a facial image. After acquiring a facial image for each frame, the averaging processor 102 sets, based on previously prepared facial shape data, a plurality of landmarks a indicating the contour positions of the cheeks, b indicating the positions of the lips, c indicating the positions of the eyes, d indicating the positions of the eyebrows, and e indicating the positions of the nose, as shown in FIG. 5 . The averaging processor 102 then determines the nose region f from the positions of the set landmarks a through e and extracts color data of pixels within this nose region f. Furthermore, the averaging processor 102 performs a process to set the values ​​of all pixels within the range (the range of data Dx in FIG. 6 ) excluding the values ​​near the minimum and maximum values ​​of the extracted color data of the pixels.

[0041] As a result, the averaging processing unit 102 extracts a specific color range of the nose area of ​​the face image and obtains the average of the (R-B) components already described for the extracted nose area. The averaging processing unit 102 then performs processing to detect the blood flow rate for the nose in the face. As a result, the stress detection processing in the detection unit 104 detects stress based on changes in blood flow rate.

[0042] [Examples of Stress Detection] Next, two examples of actual stress detection performed using the autonomic nervous activity detection system 100 of this embodiment are shown in FIGS. 13 and 14 . The two examples shown in FIGS. 13 and 14 show changes in the detected blood flow values ​​(average values) from the start to the end of an experimental task in which a user solves a prepared mental arithmetic problem. The vertical axis in FIGS. 13 and 14 represents the (R-B) component value of the nose region obtained by the averaging processor 102, and the horizontal axis represents the elapsed time (seconds) from the start of the experiment. Here, it is assumed that the averaging processor 102 performed averaging (smoothing) every 20 seconds. In FIGS. 13 and 14 , a larger detected (R-B) component value indicates a weaker level of stress, while a smaller detected (R-B) component value indicates a stronger level of stress.

[0043] The example shown in Figure 13 shows a situation in which the subject made a mistake in a mental arithmetic problem at timing P11, and then realized the mistake immediately afterward at timing P12. He also made another mistake in the mental arithmetic problem at timing P13. The detected value of the (R-B) component is high for a while after the start of the experiment, but drops sharply after timing P12, when the subject realized the mistake. Therefore, this detected value indicates that the subject was under intense stress from timing P12 onward.

[0044] Furthermore, after the detected value decreases at time P12, it is maintained at this decreased state for a certain period of time, and then after time P13, it decreases again, indicating that stress is increasing. After that (700 seconds), when the mental calculation ends, the detected value rises, indicating that stress has been relieved.

[0045] The example shown in Figure 14 is a case where the subject did not realize they had made a mistake until they finished the mental arithmetic problem. In this example, mistakes were made in the mental arithmetic problem at times P21, P22, P23, P24, P25, and P26. However, because the subject was unaware of the mistakes, there was no correlation between the mistakes and stress. In the example shown in Figure 14, the change in the detection value shows that stress increases at the start of the mental arithmetic problem, then gradually increases over time, decreasing stress. Then, after a certain amount of time (around 500 seconds) has passed, the detection value decreases in the latter half of the mental arithmetic task, indicating that stress due to fatigue is increasing.

[0046] In the example of Figure 14, when the mental calculation was completed at a later timing (700 seconds), the detected value rose, indicating that the subject was relieved from stress. However, in the example of Figure 14, it can be seen that the degree of stress reduction after the mental calculation was completed was greater than in the example of Figure 13.

[0047] As described above, the autonomic nervous activity detection system 100 of this embodiment can accurately detect autonomic nervous activity, such as stress on a subject. In this case, the image to be detected is image data obtained from primary color signals captured by a conventional camera 1. Therefore, specialized equipment, such as a far-infrared camera, that was previously required is no longer necessary, resulting in the effect of enabling accurate detection of autonomic nervous activity at low cost using existing systems. Furthermore, by using the (R-B) component, which clearly indicates blood flow, as the component to be detected, and by using a region of the (R-B) component that is limited to a range that is not affected by external light or shadows, it is possible to detect autonomic nervous activity extremely accurately.

[0048] Furthermore, in the above description, the system of this example has been described as an autonomic nervous activity detection system 100 that detects autonomic nervous activity. Alternatively, a biological detection system may be used in which, instead of detecting autonomic nervous activity, the detection unit 104 detects blood flow, and the output unit 105 outputs the detected blood flow state (biological state). A biological detection system that outputs the blood flow state (biological state) is similar to the autonomic nervous activity detection system 100 in that it utilizes the (R-B) component, which is a component that clearly indicates blood flow. This provides the distinct advantage of being able to detect and output the blood flow state more clearly than conventional methods. Furthermore, the use of an image of the nose to detect the blood flow state is merely an example, and blood flow may be detected and analyzed using similar processing from images of other biological parts besides the nose.

[0049] Furthermore, while the autonomic nervous activity detection system 100 shown in FIG. 1 is an example configured as a dedicated system for detecting autonomic nervous activity and blood flow, the autonomic nervous activity detection system 100 of this example can also be configured, for example, with a computer as shown in FIG. 3 . Therefore, by incorporating a program for operating the autonomic nervous activity detection system 100 of this example into an information processing device such as a computer or smartphone, stress detection processing can be performed in parallel with various processes being executed by the computer or smartphone, etc., making it possible to perform stress evaluations on subjects such as workers at any time. The program for operating the autonomic nervous activity detection system 100 of this example can be created by executing the processes described in the flowchart of FIG. 3 as procedures. Furthermore, this program can be stored on various types of storage media, such as memories, IC cards, SD cards, and optical discs.

[0050] 1...camera, 2...display, 100...autonomic nervous activity detection system, 100a...CPU, 100b...main memory unit, 100c...non-volatile storage, 100d...network interface, 100e...image acquisition unit, 100e...image input unit, 100f...output unit, 100g...operation unit, 101...RGB component acquisition unit, 102...averaging processing unit, 103...variation acquisition unit, 104...detection unit, 105...output unit

Claims

1. an averaging processing unit that acquires moving image data of a facial image obtained by photographing the face of a subject, and averages, for a specific region of the acquired moving image data, a difference component between a red component and a blue component among primary color components in the region at regular intervals across all target pixels in the region; a fluctuation acquisition unit that acquires the fluctuation of the difference component averaged by the averaging processing unit; a detection unit that detects blood flow or activity based on the blood flow based on the amount of variation obtained by the variation acquisition unit, The area in which the averaging processing unit performs averaging is an area obtained by excluding an area near the minimum value and an area near the maximum value of all color components from the area in which the difference component is obtained. Biodetection system.

2. Furthermore, the activity based on the blood flow is the autonomic nervous activity of the subject. The biological detection system according to claim 1 .

3. The section near the minimum value of all color components is a section in which the distribution of the difference component with respect to the light amount of all primary color components in all pixels within the specific region has a first gradient inclined from the minimum value of the light amount of all primary color components, The section near the maximum value of all the color components is a section in which the distribution of the difference components with respect to the light amounts of all the primary color components in all the pixels in the specific region has a second gradient inclined toward the maximum value of the light amounts of all the primary color components. The biological detection system according to claim 1 .

4. The specific region of the face image of the video image data acquired by the averaging processing unit is the nose region of the face. The biological detection system according to any one of claims 1 to 3.

5. A living body detection method in which an information processing device performs arithmetic processing based on moving image data of a facial image obtained by capturing a face of a subject, and detects blood flow or activity based on the blood flow of the subject, The arithmetic processing performed by the information processing device includes: an averaging process for averaging the difference components between red and blue components of the primary color components in a specific region of video image data capturing the face of a subject over all target pixels in the region at regular intervals; a fluctuation acquisition process for obtaining a fluctuation of the difference component averaged in the averaging process; a detection process for detecting blood flow or activity based on the blood flow based on the amount of variation obtained in the variation acquisition process; Including, The area to be averaged in the averaging process is an area obtained by excluding an area near the minimum value and an area near the maximum value of all color components from the area where the difference component is obtained. Liveness detection methods.

6. A program for detecting a blood flow or an activity based on the blood flow of a subject based on image data of the face of the subject, an averaging step for averaging, for a specific region of the image data, the difference component between the red component and the blue component of the primary color components in the region over all pixels at regular intervals; a fluctuation obtaining step for obtaining the fluctuation of the difference component averaged in the averaging step; a detection step of detecting a blood flow or an activity based on the blood flow based on the amount of variation obtained in the variation acquisition step; is a program that is implemented and executed on a computer, The area to be averaged in the averaging procedure is an area obtained by excluding an area near the minimum value and an area near the maximum value of all color components from the area where the difference component is obtained. program.