CN114580464A - Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis - Google Patents
Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis Download PDFInfo
- Publication number
- CN114580464A CN114580464A CN202210126540.9A CN202210126540A CN114580464A CN 114580464 A CN114580464 A CN 114580464A CN 202210126540 A CN202210126540 A CN 202210126540A CN 114580464 A CN114580464 A CN 114580464A
- Authority
- CN
- China
- Prior art keywords
- bvp
- component
- channel
- vmd
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 45
- 238000012880 independent component analysis Methods 0.000 title claims abstract description 31
- 238000009531 respiratory rate measurement Methods 0.000 title 1
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 47
- 230000003595 spectral effect Effects 0.000 claims abstract description 11
- 230000001427 coherent effect Effects 0.000 claims description 11
- 238000010183 spectrum analysis Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000000691 measurement method Methods 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 6
- 238000001228 spectrum Methods 0.000 claims description 5
- 238000012935 Averaging Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 3
- 210000001061 forehead Anatomy 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000029058 respiratory gaseous exchange Effects 0.000 claims description 3
- 230000001815 facial effect Effects 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims 1
- 230000036387 respiratory rate Effects 0.000 abstract description 45
- 238000012163 sequencing technique Methods 0.000 abstract description 5
- 230000017531 blood circulation Effects 0.000 abstract description 2
- 238000000926 separation method Methods 0.000 abstract 1
- 238000000605 extraction Methods 0.000 description 14
- 230000000694 effects Effects 0.000 description 8
- 238000011160 research Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013186 photoplethysmography Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000241 respiratory effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2134—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on separation criteria, e.g. independent component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Cardiology (AREA)
- Public Health (AREA)
- Heart & Thoracic Surgery (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Pulmonology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Veterinary Medicine (AREA)
- Bioinformatics & Cheminformatics (AREA)
Abstract
The invention discloses a human Heart Rate Variability (HRV) and Respiratory Rate (RR) measuring method based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA). Then, the G-channel signal is subjected to 4-channel decomposition by using a VMD algorithm, and a reference signal of a blood flow pulse (BVP) is obtained based on a component having the largest spectral peak among the decomposed 4-channel components. Based on the reference signal, a BVP source signal is separated from the RGB observation signal by using a cICA algorithm, a VMD algorithm is used for carrying out 4-channel decomposition on the BVP source signal, a high-quality pulse wave component is extracted from the separated 4-channel component, and HRV parameters and RR are further obtained. The method can avoid the inherent source sequencing fuzzy problem in the traditional blind source separation/independent component analysis algorithm, has better noise interference resistance, and has better application prospect in the field.
Description
Technical Field
The invention relates to the field of extraction research of face video physiological parameters based on an imaging type photoplethysmography (IPPG) technology, comprising the extraction of Heart Rate Variability (HRV) parameters and Respiratory Rate (RR), in particular to a Variational Modal Decomposition (VMD) algorithm and a constraint independent component analysis (cICA) algorithm.
Background
Heart Rate Variability (HRV) and Respiratory Rate (RR) are important clinical physiological parameters of the human body. Exploring a non-contact measurement method of HRV and RR becomes one of the hot spots of current biomedical engineering and instrument field research.
The imaging type photoplethysmography (IPPG) is a biomedical signal processing technology which analyzes video data of a human body surface sensitive area through an intelligent information processing algorithm and further extracts physiological parameters such as Heart Rate (HR), HRV and RR, and has the advantages of non-contact, convenience in operation, wide application prospect and the like.
Since video data in IPPG technology usually depends on the visible light environment of the human body surface, the generated RGB observation signals and the obtained BVP component are usually interfered by complex background noise, and this defect further affects the extraction of various physiological parameters.
Many of the patents or literature publications have disclosed circumventing or employing different algorithms to overcome this problem. For example, the Chinese patent publication No. CN113657345A was published on publication No. 2021-11-16, the Chinese patent publication No. CN112237421A was published on publication No. 2021-01-19, and the Chinese patent publication No. CN107616795A was published on publication No. 2018-01-23.
In the video-based physiological parameter extraction schemes, different pattern recognition algorithms are tried to extract target parameters, such as HRV characteristic parameter extraction by using algorithms such as adaptive threshold skin detection, LSTM convolutional network and the like, and a respiratory signal sequence is extracted from a chest video by using a video motion amplification technology.
In the existing IPPG application studies, Independent Component Analysis (ICA) method is an important idea for its superiority in extracting blood flow pulse (BVP) source signals. The extracted BVP source signal based on ICA can be used for accurately extracting HR data, and the research method is greatly shown in relevant published documents and patents.
However, the noise with different intensities remained in the BVP source signal makes it not effectively suitable for extracting physiological parameters such as HRV and RR, which have high requirements on BVP waveform quality. In recent years, a new adaptive signal processing method, variational modal decomposition, has attracted attention because of its good application effect in non-stationary and non-linear signal processing.
Through a large amount of analysis research and tests, the inventor team of the invention discovers that the VMD method can be used for decomposing ideal pulse wave components from the RGB observation signals and the BVP source signals under different noise interference environments and further extracting target physiological parameters.
Secondly, the ICA algorithm has a source sorting fuzzy problem when separating BVP source signals, and an identification algorithm of the BVP source signals needs to be additionally designed in practical application. The BVP source discrimination in the published relevant documents is basically based on methods such as spectrum analysis, and the like, and the identification accuracy of the methods cannot achieve ideal effects in a complex noise interference environment. For example, Chinese invention patent with publication number 102499664A with publication number 2012-06-20, Chinese invention patent with publication number 103271734A with publication number 2013-09-04, and so on.
The method can effectively avoid the problem, and the algorithm directly acquires the BVP source signal by introducing a reference signal carrying partial key features of the target source signal and constraining the convergence direction of a spatial filter in the iterative calculation process, so that an additional BVP source identification algorithm can be abandoned.
Disclosure of Invention
The invention aims to provide an HRV and RR measuring method based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA), which is used for synchronously extracting HRV parameters and RR from a human face video, avoiding the inherent BVP source sequencing fuzzy problem of the traditional method and well overcoming the troublesome BVP source signal noise residue problem in physiological parameter extraction.
To this end, in one aspect, the present invention provides a method for measuring HRV and RR of a human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (ica), comprising: s100, carrying out pixel coherent average operation on human face video data, and converting the human face video data into RGB observation signals; s101, preprocessing the RGB observation signals to obtain standardized observation signals for subsequent analysis; s102, performing 4-channel decomposition on the G-channel signal by using a VMD algorithm, and generating a BVP reference signal on the basis of the decomposed component with the maximum spectral peak value in the 4-channel component; s103, separating a BVP source signal from the RGB observation signal by using a cICA algorithm based on the BVP reference signal; s104, performing 4-channel decomposition on the BVP source signal by using a VMD algorithm, and extracting a high-quality pulse wave component from the decomposed 4-channel component; and S105, obtaining HRV parameters based on the high-quality pulse wave components: low frequency component power (LF), high frequency component power (HF), power ratio of low frequency component to high frequency component (LF/HF), and RR.
According to another aspect of the present invention, there is provided a human HRV and RR measurement apparatus based on Variational Modal Decomposition (VMD) and constraint independent component analysis (ica), comprising: s100, carrying out pixel coherent average operation on human face video data to convert the human face video data into a first program module of RGB observation signals; s101, preprocessing the RGB observation signals to obtain standardized observation signals for a second program module of subsequent analysis; s102, a third program module for performing 4-channel decomposition on the G-channel signal by using a VMD algorithm and solving a BVP reference signal on the basis of the decomposed component with the maximum spectral peak value in the 4-channel component; s103, based on the BVP reference signal, using a cICA algorithm to separate a BVP source signal from the RGB observation signal; s104, a program module V for performing 4-channel decomposition on the BVP source signal by using a VMD algorithm and extracting high-quality pulse wave components from the decomposed 4-channel components; and S105, a program module VI for obtaining the HRV parameters and the RR based on the high-quality pulse wave components, wherein the HRV parameters are as follows: low frequency component power (LF), high frequency component power (HF), and a power ratio of the low frequency component to the high frequency component (LF/HF).
The present invention also provides a computer readable storage medium on which a program is stored which when executed performs the steps of the method for human HRV and RR measurement based on variational modal decomposition and constraint independent component analysis according to the above described method.
The invention also provides a computer device comprising a processor and a memory, characterized in that the memory has stored thereon a program which, when executed on the processor, carries out the steps of the method for measuring HRV and RR of a human body based on variational modal decomposition and constrained independent component analysis according to the above described method.
Compared with the prior art, the human Heart Rate Variability (HRV) and Respiratory Rate (RR) measuring method based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) has the following characteristics.
1. The invention overcomes the problem of the troublesome BVP source signal noise residue in the traditional research method.
Based on a great amount of analysis and test in the prior art, the VMD, the novel adaptive signal processing method, is applied to the processing of RGB observation signals and BVP source signals, can decompose ideal pulse wave components, and is further applied to the extraction of HRV parameters and RR. Compared with the prior art, the pulse wave component with high-quality waveform can be obtained from the BVP source signal by using the VMD method, and the problem of noise residue in the BVP source signal which always troubles physiological parameter extraction is solved. In addition, the BVP reference signal acquired based on the G channel and the VMD method can carry accurate frequency and phase information, and the robustness of the cICA algorithm is improved. The schemes in the invention are not disclosed in relevant research.
2. The invention effectively solves the problem of fuzzy sequencing of the BVP source in the traditional independent component analysis scheme.
Compared with the prior art, the method can effectively avoid the BVP source signal sequencing fuzzy problem by adopting a constrained independent component analysis (cICA) method, and directly obtains the BVP source signal by constraining the convergence direction of a spatial filter in the iterative computation process by introducing the reference signal carrying partial key characteristics of the BVP source signal, thereby saving an additional BVP source identification algorithm. Under the condition of noise interference, the robustness of a physiological parameter detection algorithm can be greatly improved by the CICA-based BVP source signal extraction method.
3. The invention has great application potential.
A non-contact physiological parameter measurement method based on a facial video has gradually become a hotspot in research and application in the field of physiological parameter detection. But besides traditional heart rate detection, other physiological parameter extraction studies and applications are still in the exploration phase. The related method provided by the invention better overcomes the problems troubled by the published patents and documents, and realizes the effective measurement of HRV parameters and RR. The method achieves more than 90% of accuracy rate in multiple comparison tests with the traditional contact type physiological parameter detector, and the extracted high-quality pulse wave components can be further expanded and applied in the field of non-contact type physiological parameter detection, so that the method has good application potential.
Therefore, the HRV parameter and RR measuring method based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) provided by the invention has the advantages of stronger anti-noise interference capability, high measuring accuracy, large application potential and the like.
In addition to the objects, features and advantages described above, other objects, features and advantages of the present invention are also provided. The present invention will be described in further detail below with reference to the drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic flow chart of a method for measuring Heart Rate Variability (HRV) and Respiratory Rate (RR) according to the present invention;
FIG. 2 is a block diagram showing a detailed implementation of the method of the present invention;
FIG. 3 is a graph of the pre-processed effect of RGB observation signals generated based on selected video data samples;
FIG. 4 is a diagram of the effect of 4-channel VMD decomposition of a G-channel observation signal;
FIG. 5 is a diagram showing the effect of a BVP reference signal generated based on a G channel VMD-1 component and a BVP source signal output by a cICA algorithm;
FIG. 6 is a diagram of the effect of 4-channel VMD decomposition of a BVP source signal;
fig. 7 is a graph showing the effect of HRV parameter and RR extraction based on the high-quality pulse wave component in the BVP source signal.
In order that the present invention may be more readily understood, a more particular description of the invention briefly described above will be rendered by reference to specific embodiments that are illustrated in the appended drawings.
Detailed Description
The human HRV and RR measuring method of the invention performs pixel coherent average operation on human face video data and converts the human face video data into RGB observation signals. And preprocessing the RGB observation signal to obtain a standardized observation signal, performing 4-channel decomposition on the G-channel signal by using a VMD algorithm, and solving the BVP reference signal on the basis of the decomposed component with the maximum spectrum peak value in the 4-channel component. Further, based on the reference signal, a BVP source signal is separated from the RGB observation signal using the ica algorithm, and a VMD algorithm is used to perform 4-channel decomposition on the BVP source signal, thereby extracting a high-quality pulse wave component from the decomposed 4-channel component. And finally, obtaining HRV parameters based on the high-quality pulse wave components: low frequency component power (LF), high frequency component power (HF), power ratio of low frequency component to high frequency component (LF/HF), and RR.
Compared with the non-contact HRV parameter or RR extraction scheme provided by the disclosed method, the method can avoid the problem of BVP source sequencing ambiguity in the scheme based on the traditional ICA algorithm, has better anti-noise interference performance, and has better application prospect in the field.
The following describes the measurement method of human Heart Rate Variability (HRV) and Respiratory Rate (RR) according to the present invention.
Referring to fig. 1, the HRV and RR measurement method provided by the present invention includes the following steps:
s100, carrying out pixel coherent average operation on human face video data, and converting the human face video data into RGB observation signals;
s101, preprocessing the RGB observation signals to obtain standardized observation signals for subsequent analysis;
s102, performing 4-channel decomposition on the G-channel signal by using a VMD algorithm, and generating a BVP reference signal on the basis of the decomposed component with the largest spectral peak value in the 4-channel component;
s103, separating a BVP source signal from the RGB observation signal by using a cICA algorithm based on the BVP reference signal;
s104, performing 4-channel decomposition on the BVP source signal by using a VMD algorithm, and extracting a high-quality pulse wave component from the decomposed 4-channel component;
s105, obtaining HRV parameters based on the high-quality pulse wave components: low frequency component power (LF), high frequency component power (HF), power ratio of low frequency component to high frequency component (LF/HF), and RR.
Referring to the system implementation block diagram shown in fig. 2, the detailed implementation steps of the HRV and RR measurement method provided by the present invention are as follows:
1. the method comprises the steps of collecting face video data of a subject, wherein the used equipment is a common RGB camera (the sampling rate is 30Hz), and the subject is located at a distance of about 0.3-1 m in front of the camera.
2. And (3) calculating the face video data by using a coherent averaging method, performing spatial coherent averaging calculation on RGB pixel values of a selected sensitive area in each frame of image in the video, and converting the frame sequence into an RGB observation signal sequence, wherein the selected sensitive area is a forehead area.
3. And carrying out cubic spline interpolation on the generated RGB observation signal, and increasing the sampling rate of the signal to 300Hz so as to improve the time precision of a peak in the follow-up heartbeat interval (IBI) statistics.
4. And carrying out mean value removing, normalization and band-pass filtering denoising operations on the RGB observation signals to obtain standardized observation signals for subsequent analysis. Wherein the cut-off frequency of the band-pass filter is 0.5-4 Hz;
5. the G-channel signal is decomposed into 4 channels by using a VMD algorithm, and a component having the largest spectral peak is extracted from the decomposed 4 channels.
6. Extracting the Fourier series corresponding to the maximum spectral peak in the component, performing inverse Fourier transform on the Fourier series by using an Euler formula, and taking the real part of the complex signal obtained by inverse transform as a BVP source signal reference signal (the reference signal carries the frequency and phase information of the BVP source signal).
7. Based on the BVP reference signal, a BVP source signal is separated from the RGB observation signal using the cICA algorithm. The cICA algorithm adopts an objective function based on the maximum negative entropy, the threshold of the difference between the estimation signal and the reference signal of the BVP source signal in the iterative operation process is set to be 1.5, the learning step length is set to be 0.15, and the maximum iteration number is 400.
8. And (3) performing 4-channel decomposition on the BVP source signal by using a VMD algorithm, and taking the component with the maximum spectral peak value in the decomposed 4-channel components as a high-quality pulse wave component.
9. The time interval between every two heartbeat wave peaks, namely IBI, is calculated based on the high-quality pulse wave components, and the intermediate value of the time point of the appearance of the adjacent wave peaks is used as the time coordinate of the IBI.
10. And carrying out non-uniform sampling spectrum analysis on the IBI sequence and the corresponding time coordinate sequence by adopting an LS spectrum analysis method.
11. Based on the LS spectrum analysis result, the sum of powers in the frequency range of 0.04-0.15Hz (LF), the sum of powers in the frequency range of 0.15-0.4Hz (HF), and the power ratio of the low frequency component to the high frequency component (LF/HF) in the frequency spectrum are calculated.
12. Based on the LS spectrum analysis result, extracting a frequency point corresponding to the maximum peak within the frequency range of 0.15-0.4Hz, namely the Respiratory Rate (RR).
Referring to fig. 3, without loss of generality, in this embodiment, a segment of video data of a common subject is selected, a segment of RGB observation signals is generated by a pixel coherent averaging method based on a forehead region, and the preprocessing operation is performed.
Referring to fig. 4, which shows the 4-channel VMD decomposition of the RGB observation signal shown in fig. 3 in the present embodiment, the component containing more obvious pulse wave information is better decomposed as shown by the VMD-1 component. The component carries more complete key information such as BVP source signal frequency and phase, and can be used for generating a BVP reference signal.
Referring to fig. 5, a BVP reference signal generated from the VMD-1 component shown in fig. 4 in this embodiment is shown, and a band-constrained independent component analysis (cia) operation is performed on the RGB observation signal based on the reference signal, so as to successfully separate out a BVP source signal without additionally designing a BVP source identification method.
Referring to fig. 6, there is shown 4-channel VMD decomposition of the BVP source signal extracted using ica described in fig. 5 in the present embodiment. From the decomposition result, the component carrying the good quality pulse wave component is well decomposed, as shown by the VMD-1 component, which carries the more ideal human pulse wave information for the extraction of HRV parameters and RR.
Referring to fig. 7, an example of the details of the extraction and the measurement effect of the HRV parameters and RR based on the high-quality pulse wave components described in fig. 6 in the present embodiment is shown.
First, the time interval between the heartbeat peaks, i.e. the IBI, is calculated from the pulse wave components, and the median of the time points at which the adjacent peaks appear is taken as the time coordinate of the IBI. Further, the obtained IBI sequence and the corresponding time coordinate sequence are subjected to non-uniform sampling spectrum analysis by an LS spectrum analysis method. And finally, calculating the low-frequency component power (LF), the high-frequency component power (HF), the power ratio (LF/HF) of the low-frequency component and the high-frequency component in the frequency spectrum, and a frequency point corresponding to the maximum peak in the high-frequency band, namely RR.
The invention also provides a human body HRV and RR measuring device based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA), which comprises a program module I to a program module VI.
The program module I is used for carrying out pixel coherent average operation on the human face video data so as to convert the human face video data into RGB observation signals; the program module II is used for carrying out preprocessing operation on the RGB observation signals so as to obtain standardized observation signals for subsequent analysis; the third program module is used for carrying out 4-channel decomposition on the G-channel signal by using a VMD algorithm, and solving a BVP reference signal on the basis of the decomposed component with the maximum spectral peak value in the 4-channel component; the program module IV is used for separating a BVP source signal from the RGB observation signal by using a cICA algorithm based on the BVP reference signal; the program module V is used for carrying out 4-channel decomposition on the BVP source signal by applying a VMD algorithm and extracting a high-quality pulse wave component from the decomposed 4-channel component; and a program module six for calculating HRV parameters and RR based on the high-quality pulse wave component, wherein the HRV parameters: low frequency component power (LF), high frequency component power (HF), and a power ratio of the low frequency component to the high frequency component (LF/HF).
The HRV and RR measurement method of the human body according to the present invention is preferably integrated in an electronic device in the form of a computer processing program, and the electronic device may be a server or a terminal.
The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, Network service, cloud communication, middleware service, domain name service, security service, Network acceleration service (CDN), big data and an artificial intelligence platform.
The terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart sound box, a smart watch, and the like. The terminal and the server may be directly or indirectly connected by wired or wireless communication.
The computer-readable storage medium may include: a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, etc. which stores the HRV and RR measurement programs for human body, and the HRV and RR measurement programs are used to implement the steps of the HRV and RR measurement methods for human body when executed.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A human HRV and RR measurement method based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA), comprising:
s100, carrying out pixel coherent average operation on human face video data, and converting the human face video data into RGB observation signals;
s101, preprocessing the RGB observation signals to obtain standardized observation signals for subsequent analysis;
s102, performing 4-channel decomposition on the G-channel signal by using a VMD algorithm, and solving a BVP reference signal on the basis of the decomposed component with the maximum spectral peak value in the 4-channel component;
s103, separating a BVP source signal from the RGB observation signal by using a cICA algorithm based on the reference signal;
s104, performing 4-channel decomposition on the BVP source signal by using a VMD algorithm, and extracting a high-quality pulse wave component from the decomposed 4-channel component;
s105, obtaining HRV parameters based on the high-quality pulse wave components: low frequency component power (LF), high frequency component power (HF), power ratio of low frequency component to high frequency component (LF/HF), and Respiration Rate (RR).
2. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the step S100 of generating RGB observed signals comprises: calculating the RGB pixel value of the selected sensitive area in each frame of image in the facial video by using a coherent averaging method, and further converting the frame sequence into an RGB observation signal sequence, wherein the selected sensitive area is a forehead area.
3. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the preprocessing operation of RGB observed signals in step S101 comprises:
step 11: carrying out cubic spline interpolation on the RGB observation signals, and increasing the sampling rate from about 30Hz to 300 Hz;
step 12: carrying out mean value removing and normalization operation on the RGB observation signals;
step 13: and performing band-pass filtering and denoising on the RGB observation signals, wherein the cut-off frequency of a filter is 0.5-4 Hz.
4. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the step S102 of obtaining BVP reference signal comprises:
step 21: performing fast Fourier transform and spectrum analysis on 4-channel components decomposed by the VMD algorithm;
step 22: extracting the component with the maximum spectrum peak value in the 4-channel components;
step 23: extracting a Fourier series corresponding to a maximum spectral peak in the component;
step 24: the Fourier series is subjected to inverse Fourier transform by using an Euler formula, and a real part of a complex signal obtained by inverse transformation is taken out to be used as a BVP reference signal. The reference signal carries frequency and phase information of the BVP source signal.
5. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the cICA algorithm used in step S103 employs an objective function based on negative entropy maximization; in addition, in the iterative operation process of extracting the BVP source signal by the cICA algorithm, the threshold of the difference between the estimated signal and the reference signal of the BVP source signal is set to be 1.5, the learning step length is set to be 0.15, and the maximum iteration number is 400.
6. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the extracted high-quality pulse wave component in step S104 is the component with the largest spectral peak in 4-channel components decomposed from BVP source signal.
7. The method for measuring HRV and RR of human body based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cICA) as claimed in claim 1, wherein the step S105 of obtaining HRV parameters and RR based on the high-quality pulse wave component comprises:
step 31: counting time intervals among all wave crests in the high-quality pulse wave components, namely heartbeat intervals (IBI), and taking the intermediate value of the time points of the appearance of the adjacent wave crests as the time coordinate of the IBI;
step 32: carrying out non-uniform sampling spectrum analysis on the IBI sequence and the corresponding time coordinate sequence by adopting a Lomb-Scargle spectrum analysis method;
step 33: calculating the sum of the power in the frequency range of 0.04-0.15Hz as the low-frequency component power (LF) and the sum of the power in the frequency range of 0.15-0.4Hz as the high-frequency component power (HF); then, the power ratio (LF/HF) of the low-frequency component and the high-frequency component is obtained;
step 34: and (3) calculating a frequency point corresponding to the maximum peak in the frequency band of 0.15-0.4Hz, namely the Respiration Rate (RR).
8. A human HRV and RR measurement device based on Variational Modal Decomposition (VMD) and constraint independent component analysis (cia), comprising:
the program module I is used for carrying out pixel coherent average operation on the human face video data so as to convert the human face video data into RGB observation signals;
the program module II is used for carrying out preprocessing operation on the RGB observation signals so as to obtain standardized observation signals for subsequent analysis;
the third program module is used for carrying out 4-channel decomposition on the G-channel signal by using a VMD algorithm and solving a BVP reference signal on the basis of the decomposed component with the maximum spectrum peak value in the 4-channel components;
the program module IV is used for separating a BVP source signal from the RGB observation signal by using a cICA algorithm based on the BVP reference signal;
the program module V is used for carrying out 4-channel decomposition on the BVP source signal by using a VMD algorithm and extracting a high-quality pulse wave component from the decomposed 4-channel component;
the program module six is used for solving HRV parameters and RR based on the high-quality pulse wave components, wherein the HRV parameters are as follows: low frequency component power (LF), high frequency component power (HF), and a power ratio of the low frequency component to the high frequency component (LF/HF).
9. A computer readable storage medium on which a program is stored, which when executed performs the steps of the method for human HRV and RR measurement based on variational modal decomposition and constraint-independent component analysis of any of claims 1 to 7.
10. A computer device comprising a processor and a memory, characterized in that the memory has stored thereon a program which, when executed on the processor, carries out the steps of the method for HRV and RR measurement of a human body based on variational modal decomposition and constraint independent component analysis according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210126540.9A CN114580464A (en) | 2022-02-10 | 2022-02-10 | Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210126540.9A CN114580464A (en) | 2022-02-10 | 2022-02-10 | Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114580464A true CN114580464A (en) | 2022-06-03 |
Family
ID=81770638
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210126540.9A Pending CN114580464A (en) | 2022-02-10 | 2022-02-10 | Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114580464A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116152716A (en) * | 2023-02-24 | 2023-05-23 | 上海理工大学 | Identification method for lost mode in binocular vision dynamics mode parameter identification |
-
2022
- 2022-02-10 CN CN202210126540.9A patent/CN114580464A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116152716A (en) * | 2023-02-24 | 2023-05-23 | 上海理工大学 | Identification method for lost mode in binocular vision dynamics mode parameter identification |
| CN116152716B (en) * | 2023-02-24 | 2023-12-08 | 上海理工大学 | Identification method for lost mode in binocular vision dynamics mode parameter identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Automatic recognition of arrhythmia based on principal component analysis network and linear support vector machine | |
| CN110399857B (en) | Electroencephalogram emotion recognition method based on graph convolution neural network | |
| Zhao et al. | ECG authentication system design incorporating a convolutional neural network and generalized S-Transformation | |
| CN107529646B (en) | Non-contact heart rate measurement method and device based on Euler image amplification | |
| Wang et al. | Arrhythmia classification algorithm based on multi-head self-attention mechanism | |
| Chawla | A comparative analysis of principal component and independent component techniques for electrocardiograms | |
| Li | Wavelets for electrocardiogram: overview and taxonomy | |
| CN111310570B (en) | Electroencephalogram signal emotion recognition method and system based on VMD and WPD | |
| CN112120694A (en) | Motor imagery electroencephalogram signal classification method based on neural network | |
| CN111184509A (en) | Emotion-induced electroencephalogram signal classification method based on transfer entropy | |
| CN112890827B (en) | Electroencephalogram identification method and system based on graph convolution and gate control circulation unit | |
| Kumar et al. | Total variation denoising based approach for R-peak detection in ECG signals | |
| CN113326801A (en) | Human body moving direction identification method based on channel state information | |
| CN114612885A (en) | Driver fatigue state detection method based on computer vision | |
| CN114580464A (en) | Human heart rate variability and respiratory rate measurement method based on variational modal decomposition and constraint independent component analysis | |
| Ji et al. | Broadband noise suppression and feature identification of ECG waveforms using mathematical morphology and embedding theorem | |
| CN111931578B (en) | Electroencephalogram identification method based on minimum spanning tree and regional double-layer network | |
| CN110321781B (en) | Signal processing method and device for non-contact measurement | |
| CN116421200A (en) | Brain electricity emotion analysis method of multi-task mixed model based on parallel training | |
| CN116250849A (en) | Electroencephalogram signal identification method based on information separator and regional convolution network | |
| CN110432899B (en) | Electroencephalogram signal identification method based on depth stacking support matrix machine | |
| Emara et al. | A Hybrid Compressive Sensing and Classification Approach for Dynamic Storage Management of Vital Biomedical Signals | |
| CN104873187A (en) | Method for iterative analysis of human body bio-electricity signal by sample entropy | |
| CN112741598A (en) | Waveform generation recognition method and computer readable medium | |
| Bhagwat et al. | Map composition framework for synthetic P morphology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |