Device for estimating cardiovascular biometric information based on image, or method therefor

Abstract

A device for estimating cardiovascular biometric information based on an image is presented according to one embodiment of the present invention. The device may comprise: a memory for storing data for estimation of the cardiovascular biometric information; and a processor for estimating cardiovascular biometric information for an object in an image obtained using the data for estimation of the cardiovascular biometric information.

Description

Device for estimating image-based cardiovascular bioinformation or method for the same

[0001] The present invention relates to an apparatus or method for estimating image-based cardiovascular bioinformation, and more specifically, to an apparatus or method configured to perform estimation of cardiovascular bioinformation for an object identified in an image and additionally output the estimated cardiovascular bioinformation.

[0002] rPPG (remote photoplethysmography) refers to a technology that measures physiological signals, such as heart rate, in a non-contact manner. This technology is based on the principle of tracking blood flow by analyzing changes in light reflected from the skin.

[0003] Specific color changes (primarily in the green channel) are analyzed in images captured by a camera through the skin surface, and these color changes occur as the amount of blood in the skin changes with the heartbeat.

[0004] Sophisticated algorithms are used on captured image data to remove noise (movement, lighting changes, etc.) and extract periodic signals corresponding to blood flow.

[0005] rPPG offers high user convenience as it does not require sensors to be attached to the skin, and allows for various applications as measurements can be taken using smartphones, webcams, or CCTVs. Furthermore, it is safe and convenient as it does not directly touch the body while measuring biosignals.

[0006] Therefore, recently, it is being used or applied in fields such as healthcare for health monitoring by measuring heart rate, blood pressure, and stress levels; fitness for analyzing heart rate and physical condition during exercise; and security monitoring and emotional state analysis by identifying stress or tension.

[0007] rPPG technology is continuously advancing and evolving into a system capable of more accurate and real-time applications when combined with artificial intelligence.

[0008] The present invention aims to provide a method to improve performance degradation in high and low heart rate regions when acquiring cardiovascular biosignals, such as heart rate, using rPPG.

[0009] Furthermore, the present invention aims to provide a method to improve performance degradation caused by size limitations of image pixels or regions when acquiring cardiovascular biosignals, such as heart rate, using rPPG for multiple objects.

[0010] The problems to be solved by the present invention are not limited to the problems to be solved above, and other problems not mentioned will be clearly understood by those skilled in the art to which the present invention belongs from the description below.

[0011] According to one embodiment of the present invention, an apparatus configured to perform image-based estimation of cardiovascular-related bio-information is proposed, and the apparatus may be configured to include: a memory configured to store data for estimating said cardiovascular bio-information; and a processor configured to estimate cardiovascular bio-information for an object within an image obtained using said data for estimating said cardiovascular bio-information.

[0012] Additionally or alternatively, the processor may be configured to estimate cardiovascular bio-information for a plurality of objects within an acquired image.

[0013] Additionally or alternatively, the processor may be configured to output cardiovascular bio-information about the object through a human-machine interface.

[0014] Additionally or alternatively, the processor may be configured to crop an image region corresponding to each of a plurality of objects and / or scale up the cropped image region.

[0015] Additionally or alternatively, the processor may be configured to perform identification of a plurality of objects and output cardiovascular bio-information for the identified objects through a human-machine interface.

[0016] Additionally or alternatively, the processor may be configured to search for information matching the plurality of objects in the user information stored in the memory.

[0017] Additionally or alternatively, the processor may be configured to transmit or output a notification message as cardiovascular biometric information for any one of the plurality of objects exceeds or deviates from a preset threshold or threshold range.

[0018] Additionally, or alternatively, the data for estimating the cardiovascular bio-information may include data trained on the RGB data of an image and the corresponding PPG data. The RGB data of the image may include data obtained by preprocessing the RGB data of the image. Furthermore, the PPG data may correspond to specific values ​​of the cardiovascular bio-information.

[0019] Additionally or alternatively, the data for estimating the cardiovascular bio-information may include data trained on virtual RGB data and corresponding virtual rPPG data for preset values ​​or ranges of cardiovascular bio-information.

[0020] According to another embodiment of the present invention, a method for estimating image-based cardiovascular bio-information is proposed, the method may include the steps of: acquiring an image; detecting an object in the acquired image and acquiring a signal for a face region of the detected object; and estimating cardiovascular bio-information of the detected object based on the acquired signal.

[0021] In addition, according to another embodiment of the present invention, a computer-readable medium is proposed that stores code configured to be executed by a computer or processor for a method for estimating image-based cardiovascular bio-information.

[0022] The above-mentioned problem-solving methods are merely some of the embodiments of the present invention, and various embodiments reflecting the technical features of the present invention can be derived and understood by those skilled in the art based on the detailed description of the present invention to be described below.

[0023] The present invention has the following technical effects.

[0024] The present invention can improve performance degradation in high heart rate and low heart rate regions when acquiring cardiovascular biosignals such as heart rate using rPPG.

[0025] In addition, the present invention can improve performance degradation caused by size limitations of image pixels or regions when acquiring cardiovascular biosignals, such as heart rate, using rPPG for multiple objects.

[0026] The effects according to the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the following detailed description of the invention.

[0027] The accompanying drawings, which are included as part of the detailed description to aid in understanding the present invention, provide embodiments of the present invention and explain the technical concept of the present invention together with the detailed description.

[0028] Figure 1 is a diagram illustrating the principle of heart rate detection using PPG.

[0029] Figure 2 is a diagram illustrating the principle of rPPG.

[0030] Figure 3 shows the normal resting heart rate distribution for women and men.

[0031] Figure 4 shows the results of heart rate estimation according to the prior art.

[0032] Figure 5 illustrates the process of amplifying an rPPG signal within a preset heart rate range according to the present invention.

[0033] Figure 6 shows a flowchart of image-based heart rate estimation according to the present invention.

[0034] Figure 7 illustrates the learning process of a model for image-based heart rate estimation according to the present invention.

[0035] Figure 8 shows the result of heart rate estimation according to the present invention.

[0036] FIG. 9 shows a flowchart of image-based heart rate estimation for multiple objects according to the present invention.

[0037] FIG. 10 illustrates a block diagram of an image-based device for estimating cardiovascular-related bio-information according to the present invention.

[0038] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Identical or similar components regardless of drawing symbols will be assigned the same reference number, and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably solely for the ease of drafting the specification and do not inherently possess distinct meanings or roles. Furthermore, in describing embodiments disclosed in this specification, if it is determined that a detailed description of related prior art could obscure the essence of the embodiments disclosed in this specification, such detailed description will be omitted. Additionally, the attached drawings are intended only to facilitate understanding of the embodiments disclosed in this specification; the technical concept disclosed in this specification is not limited by the attached drawings, and it should be understood that they include all modifications, equivalents, and substitutions that fall within the spirit and technical scope of the present invention.

[0039] Terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. These terms are used solely for the purpose of distinguishing one component from another.

[0040] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between.

[0041] A singular expression includes a plural expression unless the context clearly indicates otherwise.

[0042] In this application, terms such as “comprising” or “having” are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0043]

[0044] Figure 1 is a diagram illustrating the principle of heart rate detection using PPG (Photoplethysmography).

[0045] PPG is a non-invasive optical technology that is a method for acquiring physiological signals to measure the body's blood flow behaviors resulting from changes in blood volume. PPG is primarily based on the principle of detecting changes in reflected or transmitted light after irradiating the skin with infrared (IR) or visible light. Referring to Fig. 1, the process of acquiring an analog signal is illustrated in which a light source (LED) is irradiated toward the skin, and a photodetector (PD) detects light reflected from hemoglobin in the blood.

[0046] This utilizes the principle of detecting periodic changes in blood flow according to the heartbeat by utilizing the light absorption characteristics of blood.

[0047]

[0048] Meanwhile, although this specification refers only to heart rate as cardiovascular bio-information obtained from PPG or rPPG, the present invention may be extended to estimate the following bio-information in addition to heart rate, and the scope of the rights of the present invention should include estimating the following bio-information.

[0049] - Heart Rate Variability (HRV)

[0050] - Blood oxygen saturation (SpO₂)

[0051] - Respiratory Rate (RR)

[0052] - Pulse Wave Velocity (PWV)

[0053] - Blood Pressure (BP)

[0054] - Vascular Compliance & Volume Changes

[0055] - Stress Index

[0056] - Cardiac Output (CO)

[0057] Derived Cardiac Indices

[0058] - Macpa analysis-based indicators

[0059]

[0060] Figure 2 is a diagram illustrating the principle of rPPG. Basically, rPPG is the same as PPG, but differs in that it is non-contact. Referring to Figure 2, it can be seen that both the light source and the sensor (camera) are separated from the skin.

[0061] Similar to PPG, in addition to the component of external light reflected from the skin surface, the component absorbed and scattered within the skin and reflected to the camera (diffused reflection) is monitored, and heart rate information caused by minute changes in blood flow on the skin surface can be detected.

[0062]

[0063] Figure 3 shows the normal resting heart rate distribution for women and men.

[0064] Figure 3 (a) shows the heart rate distribution of a woman at rest, and Figure 3 (b) shows the heart rate distribution of a man at rest.

[0065] It can be confirmed that women have a lower average heart rate than men.

[0066] However, in regions where the distribution of heart rates is not extensive, such as regions where the heart rate per minute is 50 or less and regions where the heart rate per minute is 80 or more, there is a high probability that heart rate estimation based on image-based rPPG will be inaccurate. This is because rPPG signals are estimated from RGB data extracted from images, and since this utilizes an inference model through a learning process, there is insufficient data for training in the aforementioned heart rate per minute ranges.

[0067]

[0068] Figure 4 shows the results of heart rate estimation according to the prior art. Figure 4 shows the results of heart rate estimation based on rPPG of an object obtained from multiple images. In Figure 4, the x-axis represents the input of a PPG signal (data set) corresponding to a known heart rate, and the y-axis represents the inference result obtained by processing the input in an inference model, i.e., the estimated value of the heart rate.

[0069] Since the x-axis and y-axis of Figure 4 have the same dimension and scale, points distributed close to a straight line with a slope of 1 represent accurate estimation results, while points distributed far from the straight line represent errors.

[0070] Referring to Figure 4, error data is confirmed in the low heart rate range (about 50 bpm or less) and the high heart rate range (about 110 bpm or more), indicated by the rectangle.

[0071] In this invention, we propose the configuration, design, or training of an inference model to obtain better estimation results in the above heart rate range, and heart rate estimation using the same.

[0072]

[0073] Figure 5 illustrates the process of amplifying an rPPG signal within a preset heart rate range according to the present invention.

[0074] As explained earlier with reference to the estimation results shown in Figure 4, the heart rate distribution illustrated in Figure 3 results in a decrease in estimation accuracy in relatively low or high heart rate ranges. Accordingly, this proposal describes a method to improve estimation accuracy in relatively low or high heart rate ranges.

[0075] An rPPG signal is provided at the very top of FIG. 5. The rPPG signal is a reference signal and may be an rPPG signal corresponding to a preset heart rate (or preset heart rate range) (hereinafter referred to as the “reference rPPG signal”). The illustrated reference rPPG signal is expressed in the time domain, so the x-axis represents time. For example, the preset heart rate may be 60 bpm.

[0076] The reference rPPG signal may include data obtained from an image (i.e., data obtained by preprocessing RGB data of at least a portion of the image) and data in which the PPG signal is paired. For example, if the preset heart rate is 60 bpm, data obtained from an image measuring 60 bpm and the PPG signal at that time may be paired with each other.

[0077] Preprocessing of RGB data of an image may include a process of obtaining an alternating current (AC) signal in the time domain extracted by filtering from an RGB signal representing R, G, and B pixel values ​​in the time domain for a region of interest in the image (e.g., the face of an object).

[0078] In addition, signal analysis processing, such as feature extraction or noise removal, can be performed by synthesizing, processing, or reconstructing the filtered RGB signals channel by channel (i.e., R, G, and B). Signal analysis processing techniques may include ICA (Independent Component Analysis), PCA (Principal Component Analysis), POS (Plane-Orthogonal-to-Skin), CHROM (Chrominance-based Method), GREEN (Green Channel Method), PVB (Peak-to-Valley-Based Method), SSR (Spatial Subspace Rotation), LGI (Local Group Invariance), etc., but the present invention is not limited thereto.

[0079] By performing a Fourier transform on the reference rPPG signal in the time domain, heart rate information can be obtained, which corresponds to the previously mentioned preset heart rate or preset heart rate range information.

[0080] A reference rPPG signal is time-scaled in the time domain to obtain a virtual rPPG signal with a heart rate or heart rate range different from the information on a preset heart rate or preset heart rate range. More specifically, the reference rPPG signal is compressed or decompressed (or time-stretched) in the time domain to obtain a virtual rPPG signal with a heart rate or heart rate range different from the information on a preset heart rate or preset heart rate range.

[0081] By compressing the reference rPPG signal in the time domain, a virtual rPPG signal with a heart rate or heart rate range higher than a preset heart rate or preset heart rate range can be obtained. Additionally, by decompressing the reference rPPG signal in the time domain, a virtual rPPG signal with a heart rate or heart rate range lower than a preset heart rate or preset heart rate range can be obtained.

[0082] The acquired virtual rPPG signal can be processed using a window (or window function) to obtain an augmented signal or data. Data processing using a window or window function can improve the accuracy of frequency analysis by mitigating boundary effects. Window functions such as a Hann window or a Hamming window may be used, but the present invention is not limited thereto.

[0083] The process of obtaining an augmented signal (or augmented data) from a reference rPPG signal (or reference rPPG data) can be expressed mathematically as follows.

[0084]

[0085] Here, x(t) represents the signal where PPG and px are paired, and px is a signal obtained by preprocessing the image signal (or data), and

[0086] x a (t) represents the time-scaled signal of x(t), where if a > 1, the time axis is compressed and the frequency of the signal increases (i.e., a high heart rate signal), and if 0 < a < 1, the time axis is expanded and the frequency decreases (i.e., a low heart rate signal).

[0087] In addition, N represents the window size, M represents the stride, and k represents the window index, and

[0088] n represents the time index within a single window, w represents the window function (Han, Hamming, etc.), and

[0089] Sig(n) Aug represents an augmented signal or augmented data. As described below, the augmented signal or augmented data is used in the training process of a model for image-based heart rate estimation, and specifically, can be used as the correct answer for the estimation model.

[0090]

[0091] Reference rPPG and augmented data consisting of PPG, px, and heart rate information can be represented as follows.

[0092] PPGpxHeart Rate (bpm) Base / Augmented Data PPG(x1)px155 Base Data PPG(x2)px260 Base Data PPG(x3)px365 Base Data … … … … PPG(xn-1)pxn-1100 Augmented Data PPG(xn)pxn110 Augmented Data

[0093]

[0094] The augmented signal or augmented data obtained in this way can be used as training data for an image-based heart rate inference model. In other words, the obtained augmented signal or augmented data is data used for heart rate estimation.

[0095]

[0096] FIG. 6 shows a flowchart of image-based heart rate estimation according to the present invention. Image-based heart rate estimation according to the present invention may be performed by a heart rate estimation device or by a processor of a heart rate estimation device. A description of the heart rate estimation device will be provided later with reference to FIG. 10. Hereinafter, the device (10) will be described simply as performing heart rate estimation.

[0097] The device (10) can be configured to acquire an image (S610). The image can be acquired through an image sensor such as a camera.

[0098] The device (10) can be configured to detect the face region of an object from an acquired image (S620).

[0099] The device (10) can be configured to perform a face mesh to extract specific landmarks (key feature points) from a face (S630). A face mesh is a technique in computer vision that models the geometric structure of a face by extracting specific landmarks of a face with high precision, and can represent the key feature points of a face in 3D or 2D space.

[0100] The device (10) can be configured to set a region of interest from the result of the face mesh, that is, the main feature points of the face, and to mask the noisy areas (e.g., eyes, mouth, etc.) (S640).

[0101] The device (10) may be configured to perform image data processing on a masked region of interest (S650). The image data processing may include preprocessing of the RGB data of the image, as described above with reference to FIG. 5.

[0102] The device (10) may be configured to obtain an inference result by using image data processed through an inference model as input (S660). The inference model may include the augmented signal or augmented data described above, or may include a model learned using the augmented signal or augmented data.

[0103]

[0104] Figure 7 illustrates the learning process of a model for image-based heart rate estimation according to the present invention.

[0105] As shown in Fig. 7, the inference model can be trained to reconstruct the signal identically to the PPG signal by using preprocessed image data (px) and green image data as inputs and setting the PPG signal as the label (correct answer). Although px and green image data are used as inputs, other inputs may be used.

[0106] Meanwhile, the augmented signal or augmented data described with reference to Fig. 5 can be used as a label (correct answer).

[0107]

[0108] Figure 8 shows the result of heart rate estimation according to the present invention.

[0109] FIG. 8 shows the result of estimating the heart rate based on the rPPG of an object obtained from multiple images, as a result of training an inference model using the augmented signal or augmented data described in FIG. 5. In FIG. 8, the x-axis represents the input of a PPG signal (data set) corresponding to a known heart rate, and the y-axis represents the inference result obtained by processing the input in the inference model, i.e., the estimated value of the heart rate.

[0110] Since the x-axis and y-axis of Fig. 8 have the same dimension and scale, points distributed close to a straight line with a slope of 1 represent accurate estimation results, while points distributed far from the straight line represent errors.

[0111] Unlike in Figure 4, it can be seen that the error data is significantly reduced in the high heart rate range (about 110 bpm or higher).

[0112]

[0113] FIG. 9 shows a flowchart of image-based heart rate estimation for multiple objects according to the present invention. Image-based heart rate estimation according to the present invention may be performed by a heart rate estimation device or by a processor of a heart rate estimation device. A description of the heart rate estimation device will be provided later with reference to FIG. 10. Hereinafter, the device (10) will be described simply as performing heart rate estimation.

[0114] Unlike FIG. 6, FIG. 10 is for heart rate estimation for multiple objects, so S930 has been added, and S940 to S960 must be repeated. That is, S910 and S920 correspond to S610 and S620 of FIG. 6, and S970 also corresponds to S660 of FIG. 6. In addition, S940 to S960 correspond to S630 to S650 of FIG. 6, and since they are performed for each object, they differ in that they can be repeated.

[0115] In conclusion, only S930 needs to be explained, and the remaining steps should be referred to in the explanation of Fig. 6.

[0116] The device (10) may be configured to crop an area for each object in an image, and / or perform image processing on an area for each object (S930).

[0117] Cropping of the image area can be performed by distinguishing the face area of ​​each object.

[0118] Image processing of regions for each object is performed to remove noise or improve image quality when there are multiple objects, as the number of image pixels for each object is small, unlike when there is only a single object. Image processing for each object region may include image processing such as scale-up or pixel interpolation.

[0119] As pixel interpolation techniques, Nearest Neighbor Interpolation, Bilinear Interpolation, Bicubic Interpolation, Spline Interpolation, Lanczos Interpolation, Gaussian Interpolation, etc. may be used, but the present invention is not limited thereto.

[0120] Meanwhile, image processing in S930 may be performed optionally. That is, image processing in S930 may be performed when the resolution or noise condition of the area for each object may have a meaningful effect on the result of heart rate estimation. To this end, the device (10) may be configured to determine whether image processing is required for the area for each object.

[0121] The device (10) can be configured to obtain an inference result by using an image-processed (face) region image of each object as input to an inference model (S970).

[0122] Heart rate estimation for multiple objects can be utilized in multi-user facilities such as hospitals or elderly care facilities. For example, in multi-bed wards, monitoring is currently performed by having each patient wear contact-type measuring devices for heart rate, oxygen saturation, etc.; however, according to the device or method for estimating heart rate, etc. for multiple objects of the present invention, heart rate, etc. for multiple users can be estimated using an image sensor such as a camera and tracked and monitored. It is necessary to identify multiple users.

[0123]

[0124] FIG. 10 illustrates a block diagram of an image-based device for estimating cardiovascular-related bio-information according to the present invention.

[0125] The cardiovascular-related bio-information estimation device (10) may be configured to include a memory (100) and a processor (101).

[0126] The memory (100) may be configured to store data for estimating cardiovascular bio-information. The data for estimating cardiovascular bio-information may include the inference model described above.

[0127] In addition, the data for estimating cardiovascular bio-information may include data trained on the RGB data of an image and the corresponding PPG data. The RGB data of the image may include data obtained by preprocessing the RGB data of the image. Furthermore, the PPG data may correspond to specific values ​​of cardiovascular bio-information.

[0128] In addition, data for estimating cardiovascular bio-information may include data trained on virtual RGB data and corresponding virtual rPPG data for preset values ​​or ranges of cardiovascular bio-information.

[0129] For example, if the cardiovascular biometric information is heart rate, the preset range of cardiovascular biometric information may be 50 or less heart rate per minute or 80 or more heart rate per minute.

[0130] The processor (101) may be configured to estimate cardiovascular bio-information for an object in an acquired image using data for estimating the cardiovascular bio-information.

[0131] Additionally, the processor (101) may be configured to estimate cardiovascular bio-information for multiple objects within the acquired image.

[0132] The processor (101) may be configured to output cardiovascular bio-information about the object through a human-machine interface.

[0133] The processor (101) may be configured to crop an image area corresponding to each of a plurality of objects and / or scale up the cropped image area.

[0134] The processor (101) may be configured to perform identification of multiple objects and output cardiovascular bio-information for the identified objects through a human-machine interface.

[0135] The processor (101) may be configured to search for information matching the plurality of objects in the user information stored in the memory. The user information stored in the memory may include user characteristic information.

[0136] In this way, the processor (101) can identify multiple objects (or users). Through the identification of multiple objects, it is possible to track and monitor cardiovascular biometric information for each object. In addition to tracking and monitoring each object, it can be configured to output a notification via a human-machine interface or a transceiver for wired or wireless communication when the cardiovascular biometric information for a specific object exceeds or deviates from a preset threshold or threshold range.

[0137] Additionally, the heart rate estimation device (10) may be configured to additionally include a human-machine interface (HMI) (102).

[0138] The human-machine interface (HMI) (102) includes means for providing visual or auditory notifications to a user or administrator, such as a display, a speaker (buzzer), or an LED light. The notifications that the HMI (102) can provide may include not only visual or auditory notifications but also haptic notifications such as vibration, and the present invention is not limited thereto.

[0139]

[0140] Even if not described with reference to FIG. 10, the cardiovascular information estimation device (10) of the present invention may perform the operation according to the present invention as described above in FIG. 2, FIG. 5 to 7, and FIG. 9.

[0141]

[0142] In addition, as another aspect of the present invention, the operation of the above-described proposal or invention may be provided as code that can be implemented, practiced, or executed by a "computer" (a comprehensive concept including a system on chip (SoC) or (micro)processor, etc.), or as a computer-readable storage medium or computer program product that stores or contains said code, and the scope of the present invention may be extended to said code or as a computer-readable storage medium or computer program product that stores or contains said code.

[0143]

[0144] The detailed description of the preferred embodiments of the present invention disclosed above is provided to enable those skilled in the art to implement and practice the present invention. Although the present invention has been described with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the present invention as described in the following claims. Accordingly, the present invention is not intended to be limited to the embodiments shown herein, but to be given the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A device configured to perform image-based cardiovascular bio-information estimation, A memory configured to store data for estimating the above cardiovascular bio-information; and A device comprising a processor configured to estimate cardiovascular bioinformation for an object within an image obtained using data for estimating the above-mentioned cardiovascular bioinformation.

2. The device according to claim 1, wherein the processor is configured to estimate cardiovascular bio-information for a plurality of objects within the acquired image.

3. In either Paragraph 1 or Paragraph 2, the processor A device configured to output cardiovascular bio-information about the above object through a human-machine interface.

4. In paragraph 2, the processor Crop an image area corresponding to each of the above plurality of objects, and / or A device configured to scale up the cropped image area.

5. In paragraph 2, the processor A device configured to perform identification of the above-mentioned plurality of objects and output cardiovascular bio-information for the identified objects through a human-machine interface.

6. In paragraph 5, the above processor A device configured to search for information matching a plurality of objects in user information stored in the memory.

7. In paragraph 5, the above processor A device configured to transmit or output a notification message when cardiovascular biometric information for any one of the plurality of objects exceeds or deviates from a preset threshold or threshold range.

8. In paragraph 1, the data for estimating the cardiovascular bio-information is: It includes data learned from RGB data of an image and corresponding PPG (Photoplethysmography) data, and The RGB data of the above image includes data obtained by preprocessing the RGB data of the image, and The above PPG data is a device corresponding to a specific value of cardiovascular bio-information.

9. In paragraph 1, the data for estimating the cardiovascular bio-information is: A device comprising data learned from virtual RGB data for preset values ​​or ranges of cardiovascular bio-information and corresponding virtual rPPG (remote photoplethysmography) data.

10. As a method for estimating image-based cardiovascular bioinformation, Step to acquire an image; A step of detecting an object in the acquired image and acquiring a signal for the face region of the detected object; and A method comprising the step of estimating cardiovascular bio-information of the detected object based on the signal obtained above.

11. A method according to claim 10, comprising the step of estimating cardiovascular bio-information for a plurality of objects within the acquired image.

12. In either Paragraph 10 or Paragraph 11, A method comprising the step of outputting cardiovascular bio-information for the above object through a human-machine interface.

13. In Paragraph 11, A step of cropping an image area corresponding to each of the plurality of objects; and / or A method comprising the step of scaling up the cropped image area.

14. In Paragraph 11, A method comprising the step of performing identification of the plurality of objects and outputting cardiovascular bio-information for the identified objects through a human-machine interface.

15. In Paragraph 14, A method comprising the step of searching for information matching the plurality of objects in user information stored in memory.

16. A method according to claim 14, comprising the step of transmitting or outputting a notification message as cardiovascular biometric information for any one of the plurality of objects exceeds or deviates from a preset threshold or threshold range.

17. In Paragraph 10, the data for estimating the above cardiovascular bioinformation is: It includes data trained on the RGB data of an image and the corresponding PPG data, and The RGB data of the above image includes data obtained by preprocessing the RGB data of the image, and The above PPG data corresponds to a specific value of cardiovascular bioinformation, a method.

18. In Paragraph 10, the data for estimating the above cardiovascular bioinformation is: A method comprising data learned from virtual RGB data for preset values ​​or ranges of cardiovascular bio-information and corresponding virtual rPPG (remote photoplethysmography) data.

19. A computer-readable medium storing code configured to be executed by a computer or processor according to any one of paragraphs 10 through 18.

Images