A driver heart rate detection method and device based on signal quality evaluation

Abstract

The application discloses a driver heart rate detection method and device based on signal quality evaluation, and belongs to the technical field of video non-contact heart rate detection. The method comprises the following steps: collecting a face video signal of a driver, obtaining a face detection area through the face video signal; dividing the face detection area, obtaining a face pulse signal based on the face detection area, and processing the face pulse signal; performing signal quality evaluation on the processed face pulse signal; and performing heart rate detection by using the evaluated face pulse signal. By introducing the signal quality evaluation, the driver face pulse signal quality collected is evaluated in real time, the heart rate detection strategy is adjusted according to the evaluation result, the detection strategy can be adaptively adjusted under different vehicle-mounted environments, the motion interference and light interference of the heart rate detection under the vehicle-mounted environment are overcome, and the robustness of the heart rate detection is improved.

Description

Technical Field

[0001] This invention relates to the field of video non-contact heart rate detection technology, and more specifically, to a driver heart rate detection method and device based on signal quality assessment. Background Technology

[0002] Human factors are the primary cause of road traffic accidents, and continuous monitoring of driver physiological status can effectively prevent accidents. Driver monitoring systems can monitor driver behavior in real time, and heart rate changes are an effective method for assessing driver fatigue and concentration, playing a crucial role in ensuring driving safety.

[0003] Traditional heart rate measurement methods using electrocardiograms or pulse oximeters require direct contact between the equipment and the driver, causing discomfort during driving. However, with the rapid development of computer vision, video photoplethysmography (MPA) detects subtle color changes in the face caused by blood pulsation, providing a non-contact basis for measuring vital signs. Near-infrared cameras in the cockpit increase the brightness reflected from the face, allowing pulse measurement without interfering with the driver. However, the near-infrared video of the driver acquired by the vehicle-mounted camera is monochromatic, and near-infrared light has a longer wavelength than visible light, making it more susceptible to motion artifacts when penetrating skin tissue, leading to inaccuracies in heart rate measurement due to the technology itself. Furthermore, motion artifacts caused by vehicle movement, driver conversation, facial expressions, and actions such as checking rearview mirrors, along with uneven ambient light distribution and rapid, drastic changes in facial brightness, all contribute to inaccurate heart rate extraction due to environmental interference.

[0004] A search revealed Chinese patent application number 202410156801.0, published on June 7, 2024, which discloses a method and device for driver in-vehicle health checks. This method describes using an in-vehicle DMS camera to detect the driver's heart rate, heart rate variability, respiratory rate, and blood pressure when the vehicle is stationary, and generates a health report. However, this method lacks a signal quality assessment step, resulting in poor accuracy of the detected indicators. Furthermore, it primarily detects the driver's heart rate in a stationary environment, failing to address the problem of robust heart rate detection while the vehicle is in motion. Therefore, the relevant technology does not offer any technical guidance on the insufficient accuracy of driver heart rate detection in dynamic environments. Summary of the Invention

[0005] 1. Technical problems to be solved

[0006] To address the problem of insufficient accuracy in driver heart rate detection in existing vehicle-mounted DMS systems, this invention provides a driver heart rate detection method and device based on signal quality assessment. By integrating signal filtering and quality assessment methods, it achieves highly accurate driver heart rate detection.

[0007] 2. Technical Solution

[0008] The objective of this invention is achieved through the following technical solutions.

[0009] A driver's heart rate detection method based on signal quality assessment includes the following steps:

[0010] The driver's facial video signal is collected, and the facial detection area is obtained from the facial video signal;

[0011] The face detection area is divided, and the facial pulse signal is obtained based on the face detection area. The facial pulse signal is then processed.

[0012] The signal quality of the processed facial pulse signal was evaluated.

[0013] Heart rate was detected using the assessed facial pulse signal.

[0014] As a further improvement of the present invention, a face detection area is divided by a face recognition tool, and a key area in the face detection area is extracted and further divided into several sub-regions.

[0015] As a further improvement of the present invention, the original facial pulse signal is extracted from several sub-regions, and a bandpass filter is used to remove noise interference from the original facial pulse signal.

[0016] As a further improvement of the present invention, the denoised original facial pulse signal is converted into a multi-dimensional vector sequence by delay coordinate transformation.

[0017] As a further improvement of the present invention, a facial pulse signal is extracted from a multidimensional vector sequence based on an orthogonal matrix image transformation filtering method. Specifically, the method includes: in the multidimensional vector sequence, let A represent the n×3 matrix of the original pulse signal, n represent the signal window length, and 3 represent the number of delayed channels. QR decomposition is achieved through Householder reflection. Householder reflection iteratively transforms the n×3 matrix A of the original pulse signal into an upper triangular matrix R. The orthogonal transformation is accumulated into the decomposed orthogonal matrix Q. The column vectors of the decomposed orthogonal matrix Q include q1, q2, and q3. The q1, q2, and q3 form the standard orthogonal basis of the column space of the n×3 matrix A of the original pulse signal.

[0018] As a further improvement of the present invention, the noise component captured by q1 is removed, and a projection matrix P is constructed to project the original facial pulse signal onto a subspace orthogonal to q1. The projection matrix P is defined as follows:

[0019] P = I n -SS T

[0020] Among them, I n Let S = q1, S denote an n×n identity matrix. T This indicates that q1 is transposed.

[0021] As a further improvement of the present invention, after obtaining the projection matrix P, the n×3 matrix A of the original pulse signal is projected onto an orthogonal subspace to eliminate the influence of noise components, thereby obtaining the facial pulse signal; the orthogonal projection is expressed as:

[0022] Y = PA

[0023] Where Y represents the facial pulse signal of an n×3 matrix.

[0024] As a further improvement of the present invention, the processed facial pulse signal is evaluated for signal quality by analyzing the features of the facial pulse signal and calculating the comprehensive score of the features using a weighted average method.

[0025] Define n features to extract: F1, F2, ..., F n Each feature corresponds to a score S1, S2, ..., S n The overall score is calculated using a weighted average method, and the formula is as follows:

[0026] S total = w1·S1 + w2·S2 + ... + w n ·S n

[0027] Among them, S total This represents the overall score, w1, w2, w n This represents the weight corresponding to each feature.

[0028] As a further improvement of this invention, the facial pulse signal acquired in each region is Fourier transformed to the frequency domain of the heart rate frequency of each region's signal, and the heart rate of each region is calculated using the heart rate frequency. The calculation formula is as follows:

[0029]

[0030] Where t represents the signal acquisition time, i represents the i-th facial region, and HR i [t] represents the heart rate result measured in the i-th facial region, Freq iThe frequency of the facial pulse signal obtained by Fourier transform is represented by fps, where fps represents the video frame rate and N represents the total number of frames of the signal.

[0031] A driver heart rate detection device based on signal quality assessment, comprising:

[0032] The signal acquisition module acquires the driver's facial video signal and obtains the facial detection area through the facial video signal;

[0033] The signal processing module divides the face detection area, obtains the facial pulse signal based on the face detection area, and processes the facial pulse signal.

[0034] The signal quality assessment module evaluates the signal quality of the processed facial pulse signal.

[0035] The heart rate detection module uses the evaluated facial pulse signal to detect heart rate.

[0036] 3. Beneficial effects

[0037] Compared with the prior art, the advantages of this invention are:

[0038] This invention provides a driver heart rate detection method and device based on signal quality assessment. By introducing a signal quality assessment step, the quality of the acquired infrared facial signal of the driver is evaluated in real time. The heart rate detection strategy is adjusted according to the assessment results, allowing for adaptive adjustment of the detection strategy under different vehicle environments. The signal quality assessment of this invention is based on multi-level feature extraction and a real-time feedback mechanism, enabling adaptive adjustment of the detection strategy under different vehicle environments. This overcomes motion interference and illumination interference in heart rate detection under vehicle conditions, improving the robustness of heart rate detection and demonstrating strong practicality and wide applicability. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the driver's heart rate detection method according to an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram comparing the driver's heart rate detection value with the reference value in an embodiment of the present invention. Detailed Implementation

[0041] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0042] Example

[0043] like Figure 1As shown in this embodiment, a driver heart rate detection method based on signal quality assessment is provided, including the following steps: acquiring the driver's facial video signal, obtaining a facial detection area through the facial video signal; dividing the facial detection area, obtaining a facial pulse signal based on the facial detection area, and processing the facial pulse signal; assessing the signal quality of the processed facial pulse signal; and using the assessed facial pulse signal to detect the heart rate.

[0044] Specifically, in this embodiment, an in-vehicle infrared camera is used to collect infrared video of the driver's face under various conditions, including when the vehicle is stationary, during low-speed daytime driving, during low-speed nighttime driving, and when the driver's face is moving.

[0045] Specifically, a monochrome CMOS image sensor is used in the vehicle system, equipped with two 940nm wavelength near-infrared LEDs as illumination sources. The driver's face is captured in real-time via the vehicle's infotainment system, and the video is stored as a dataset. In this embodiment, the camera is positioned below the steering wheel, directly facing the driver, at an average distance of 0.3 to 0.5 meters. The video is recorded at a resolution of 1280 pixels × 800 pixels, with a frame rate of 30 frames per second and a duration exceeding one minute.

[0046] Therefore, the collected driver facial infrared videos contain data from common driving scenarios, ensuring data diversity. The driver's facial video signal is captured from the acquired facial infrared videos, and the facial detection area is obtained from the captured facial video signal.

[0047] Furthermore, using the existing Seetaface facial recognition tool, the entire face is divided into multiple detection regions, including the forehead, cheeks, and eyes. Key regions are extracted from these detection regions and further subdivided into several sub-regions. In this embodiment, the key regions include the cheek and chin regions, excluding noisy areas such as facial features and the easily obscured forehead. It should be noted that motion and lighting interference during driving may affect signal extraction quality. Sub-regions allow for separate signal extraction from each divided region, using the facial video signal from the most stable region to calculate the heart rate. In this embodiment, the key regions are divided into 42 sub-regions. Since each sub-region corresponds to a different part of the face, different sub-regions provide blood flow information with varying stability under different driving environments. Furthermore, the original pulse signal is dynamically extracted from the facial video signals within the 42 sub-regions using existing video photoplethysmography.

[0048] Furthermore, a bandpass filter is used to remove noise interference from the original facial pulse signal. Specifically, in a vehicle environment, the extracted original pulse signal contains strong noise interference due to factors such as the driver's subtle facial movements, vehicle vibrations, and changes in ambient lighting, affecting the accuracy of heart rate extraction. Since a bandpass filter allows signals within a specific frequency range to pass through while suppressing frequency components outside that range, it can also filter out strong interference. Therefore, in this embodiment, in vehicle-mounted heart rate detection, the frequency range of the bandpass filter is set between 0.7Hz and 2.5Hz, thereby removing noise interference from the original facial pulse signal.

[0049] Due to the nonlinearity of physiological signals, the time-delay embedding method allows for the extraction of complex dynamic features hidden within a single signal. The time-delay embedding theorem reconstructs a high-dimensional state space from a single observable signal, equivalent to the system's true state space. Therefore, in this embodiment, the denoised original facial pulse signal is converted into a multi-dimensional vector sequence through a time-delay coordinate transformation. Specifically, for the single-channel signal A_k[n] in the k-th key region, a time-delay coordinate transformation is used to convert it into a multi-dimensional vector sequence. A time delay and an embedding dimension are defined to construct a high-dimensional vector space, which is represented as:

[0050] X k [n] = [a k [n], a k [n+τ],a k [n+2τ],…a k [n+(m-1)τ]]

[0051] Where n and k both represent natural numbers, X k [n] represents a high-dimensional vector space, a k [n] represents the average pixel value in the key region of the nth frame, m represents the embedding dimension, and τ represents the delay time, which can be selected according to the time series characteristics.

[0052] An embedding matrix can be constructed from multiple consecutive delay vectors, and the embedding matrix is ​​represented as:

[0053]

[0054] Among them, A k Let H represent the embedding matrix, and let Q represent an n×3 orthogonal matrix, where Q = H1H2…H k H k This represents the Householder reflector, used to zero out the diagonal elements in each column of the n×3 matrix A of the original pulse signal.

[0055] Furthermore, facial pulse signals are extracted from a multidimensional vector sequence using an orthogonal matrix image transformation filtering method. Specifically, in the multidimensional vector sequence, let A represent the n×3 matrix of the original pulse signal, where n represents the signal window length and 3 represents the number of delayed channels. QR decomposition is achieved through Householder reflection. Householder reflection iteratively transforms the n×3 matrix A of the original pulse signal into an upper triangular matrix R, and the orthogonal transformation is accumulated into the decomposed orthogonal matrix Q. The decomposition process is expressed as follows:

[0056] A = QR

[0057] After three iterations (k=3), the product of the Householder reflectors produces an orthogonal matrix Q. The column vectors of the orthogonal matrix Q include q1, q2, and q3. q1, q2, and q3 form the standard orthogonal basis of the column space of the n×3 matrix A of the original pulse signal.

[0058] In driving scenarios, there are significant energy changes due to ambient lighting variations, sensor noise, and facial motion. The direction that causes the most significant change in the facial pulse signal captured by q1 typically represents noise. Therefore, after removing the noise component captured by q1, a projection matrix P is constructed to project the original facial pulse signal onto a subspace orthogonal to q1. The projection matrix P is defined as follows:

[0059] P = I n -SS T

[0060] Among them, I n Let S = q1, S denote an n×n identity matrix. T This indicates that q1 is transposed.

[0061] After obtaining the projection matrix P, the n×3 matrix A of the original pulse signal is projected onto an orthogonal subspace to eliminate the influence of noise components, thus obtaining the facial pulse signal; the orthogonal projection is expressed as:

[0062] Y = PA

[0063] Here, Y represents the facial pulse signal of an n×3 matrix, where the first column has been effectively suppressed, and the second column is taken as the filtered facial pulse signal, which retains the physiological changes related to blood volume changes while minimizing irrelevant noise.

[0064] The filtered facial pulse signal is subjected to signal quality assessment by analyzing its characteristics. In this embodiment, the characteristics of the facial pulse signal include signal-to-noise ratio, spectral characteristics, and temporal stability. A weighted average method is used to calculate the comprehensive score of the characteristics.

[0065] Define n features to extract: F1, F2, ..., F n Each feature corresponds to a score S1, S2, ..., S n The overall score is calculated using a weighted average method, and the formula is as follows:

[0066] S total = w1·S1 + w2·S2 + ... + w n ·S n

[0067] Among them, S total This represents the overall score, w1, w2, w n This represents the weight corresponding to each feature.

[0068] Then, heart rate detection is performed using the evaluated facial pulse signals. Cluster analysis is conducted on the heart rate results of several sub-regions. Based on the short-term stationarity of heart rate, results with relatively stable heart rates are selected as the heart rate results for that sub-region. The heart rate information from multiple regions is then fused to obtain the final heart rate estimate. Heart rate should be the dominant component in the clustering results. Therefore, the number of samples in each cluster is counted, and clusters with the fewest samples and those with too few data points are removed, along with obvious noise points. Specifically, the facial pulse signals acquired in each region are Fourier transformed to the frequency domain to obtain the dominant heart rate frequency of each region's signal. The heart rate for each region is then calculated using the dominant heart rate frequency, with the following formula:

[0069]

[0070] Where t represents the signal acquisition time, i represents the i-th facial region, and HR i [t] represents the heart rate result measured in the i-th facial region, Freq i The frequency of the facial pulse signal obtained by Fourier transform is represented by fps, where fps represents the video frame rate, and N represents the total number of frames in the signal. In this embodiment, the total number of frames is 512, which is sufficient to obtain the heart rate results over a continuous period of time. In this embodiment, an existing multi-region heart rate fusion algorithm based on a Gaussian mixture model is used to integrate the heart rate information from each sub-region, and to filter and output the most representative heart rate detection results.

[0071] like Figure 2 The image shows the heart rate detection results. Specifically, it illustrates the results of continuous heart rate monitoring over a 10-minute driving session, including heart rate reference values ​​recorded by a pulse oximeter and heart rate measurements calculated using the detection method provided in this embodiment. Figure 2 It can be seen that when the heart rate reference value changes, the heart rate measurement value calculated in this embodiment remains well synchronized with the heart rate reference value.

[0072] This embodiment provides a driver heart rate detection method based on signal quality assessment. By introducing a signal quality assessment method, the quality of the acquired driver infrared facial signal is evaluated in real time. The heart rate detection strategy is adjusted according to the assessment results. Low-quality signals are directly removed, and high-quality signals are filtered and processed before calculating the heart rate. The signal quality assessment is based on a multi-level feature extraction and real-time feedback mechanism, which can adaptively adjust the detection strategy in different vehicle environments, overcome motion interference and light interference in heart rate detection in vehicle environments, and improve the robustness of heart rate detection.

[0073] This embodiment also provides a driver heart rate detection device based on signal quality assessment, including a signal acquisition module, a signal processing module, a signal quality assessment module, and a heart rate detection module. The signal acquisition module acquires the driver's facial video signal and obtains a facial detection region from the facial video signal. The signal processing module divides the facial detection region, obtains a facial pulse signal based on the facial detection region, and processes the facial pulse signal. The signal quality assessment module performs signal quality assessment on the processed facial pulse signal. The heart rate detection module uses the assessed facial pulse signal to detect the heart rate. The driver heart rate detection device based on signal quality assessment provided in this embodiment can implement any of the methods described in the driver heart rate detection method based on signal quality assessment, and the specific working process of the driver heart rate detection device based on signal quality assessment can refer to the corresponding process in the embodiment of the driver heart rate detection method based on signal quality assessment. The method and device provided in this embodiment can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for example, the division of a certain module is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the interconnections or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0074] This embodiment also provides a computer device. A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned driver heart rate detection method based on signal quality assessment.

[0075] This embodiment also provides a computer-readable storage medium. A computer-readable storage medium stores a computer program, which, when executed by a processor, performs the driver heart rate detection method based on signal quality assessment described in this embodiment. The computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0076] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. The accompanying drawings are only one embodiment of the present invention, and the actual structure is not limited thereto. No reference numerals in the claims should limit the scope of the claims. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the scope of protection of the present invention. Furthermore, the word "comprising" does not exclude other elements or steps, and the word "a" preceding an element does not exclude the inclusion of "a plurality" of that element. Multiple elements stated in the product claims may also be implemented by a single element through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

Claims

1. A driver's heart rate detection method based on signal quality assessment, comprising the following steps: The driver's facial video signal is collected, and the facial detection area is obtained from the facial video signal; The face detection area is divided, and the facial pulse signal is obtained based on the face detection area. The facial pulse signal is then processed. The signal quality of the processed facial pulse signal was evaluated. Heart rate was detected using the assessed facial pulse signal. The processing of the facial pulse signal includes: dividing the face detection region using a face recognition tool, extracting key regions within the face detection region, further dividing the key regions into several sub-regions, extracting the original facial pulse signal from the sub-regions, and removing noise interference from the original facial pulse signal using a bandpass filter; converting the denoised original facial pulse signal into a multi-dimensional vector sequence through delayed coordinate transformation; and extracting the facial pulse signal from the multi-dimensional vector sequence based on an orthogonal matrix image transformation filtering method, specifically including: in the multi-dimensional vector sequence, let A represent the n×3 matrix of the original pulse signal, and n represent the signal strength. The window length is 3, where 3 represents the number of delayed channels. QR decomposition is achieved through Householder reflection. Householder reflection iteratively transforms the original n×3 matrix A of the pulse signal into an upper triangular matrix R, accumulating the orthogonal transformation into the decomposed orthogonal matrix Q. The column vectors of the decomposed orthogonal matrix Q include q1, q2, and q3, which form the orthonormal basis of the column space of the original n×3 matrix A of the pulse signal. Noise components captured by q1 are removed, and a projection matrix P is constructed to project the original facial pulse signal onto a subspace orthogonal to q1. The projection matrix P is defined as: ； Among them, I n Let S = q1, S denote an n×n identity matrix. T This indicates that q1 is transposed. After obtaining the projection matrix P, the n×3 matrix A of the original pulse signal is projected onto an orthogonal subspace to eliminate the influence of noise components, thus obtaining the facial pulse signal; the orthogonal projection is expressed as: ； Where Y represents the facial pulse signal of an n×3 matrix; The processed facial pulse signal was evaluated for signal quality by analyzing the characteristics of the facial pulse signal and calculating the comprehensive score of the characteristics using a weighted average method. Define n features to extract: F1, F2, ..., F n Each feature corresponds to a score S1, S2, ..., S n The overall score is calculated using a weighted average method, and the formula is as follows: ； Among them, S total This represents the overall score, w1, w2, w n This represents the weight corresponding to each feature; The facial pulse signal acquired in each region is Fourier transformed to obtain the dominant heart rate frequency of the signal in the frequency domain. The heart rate of each region is then calculated using the dominant heart rate frequency. The calculation formula is as follows: Where t represents the signal acquisition time, i represents the i-th facial region, and HR i [t] represents the heart rate result measured in the i-th facial region, Freq i The frequency of the facial pulse signal obtained by Fourier transform is represented by fps, where fps represents the video frame rate and N represents the total number of frames of the signal.

2. A driver heart rate detection device based on signal quality assessment, characterized in that, include: The signal acquisition module acquires the driver's facial video signal and obtains the facial detection area through the facial video signal; The signal processing module divides the face detection area, obtains the facial pulse signal based on the face detection area, and processes the facial pulse signal. The process involves using a face recognition tool to segment the face detection region and extracting key regions from these regions. These key regions are then further divided into sub-regions, from which the original facial pulse signal is extracted. A bandpass filter is used to remove noise interference from the original facial pulse signal. The denoised original facial pulse signal is then converted into a multi-dimensional vector sequence using delayed coordinate transformation. Finally, the facial pulse signal is extracted from the multi-dimensional vector sequence using an orthogonal matrix image transformation filtering method. Specifically, in the multi-dimensional vector sequence, let A represent the n×3 matrix of the original pulse signal, where n represents the signal window length, and 3 represents... The number of channels with delay is indicated. QR decomposition is achieved through Householder reflection. Householder reflection iteratively transforms the original n×3 matrix A of the pulse signal into an upper triangular matrix R, accumulating the orthogonal transformation into the decomposed orthogonal matrix Q. The column vectors of the decomposed orthogonal matrix Q include q1, q2, and q3, which form the orthonormal basis of the column space of the original n×3 matrix A of the pulse signal. Noise components captured by q1 are removed, and a projection matrix P is constructed to project the original facial pulse signal onto a subspace orthogonal to q1. The projection matrix P is defined as: ； Among them, I n Let S = q1, S denote an n×n identity matrix. T This indicates that q1 is transposed. After obtaining the projection matrix P, the n×3 matrix A of the original pulse signal is projected onto an orthogonal subspace to eliminate the influence of noise components, thus obtaining the facial pulse signal; the orthogonal projection is expressed as: ； Where Y represents the facial pulse signal of an n×3 matrix; The signal quality assessment module evaluates the signal quality of the processed facial pulse signal. Among them, the signal quality of the processed facial pulse signal is evaluated by analyzing the characteristics of the facial pulse signal and calculating the comprehensive score of the features using a weighted average method. Define n features to extract: F1, F2, ..., F n Each feature corresponds to a score S1, S2, ..., S n The overall score is calculated using a weighted average method, and the formula is as follows: ； Among them, S total This represents the overall score, w1, w2, w n This represents the weight corresponding to each feature; The heart rate detection module uses the evaluated facial pulse signal to detect heart rate. Specifically, the facial pulse signal acquired in each region is Fourier transformed to obtain the dominant heart rate frequency of the signal in the frequency domain. The heart rate of each region is then calculated using the dominant heart rate frequency. The calculation formula is as follows: ； Where t represents the signal acquisition time, i represents the i-th facial region, and HR i [t] represents the heart rate result measured in the i-th facial region, Freq i The frequency of the facial pulse signal obtained by Fourier transform is represented by fps, where fps represents the video frame rate and N represents the total number of frames of the signal.

Images