CN112869737A - Non-contact human body blood oxygen saturation detection method - Google Patents

Non-contact human body blood oxygen saturation detection method Download PDF

Info

Publication number
CN112869737A
CN112869737A CN202110136278.1A CN202110136278A CN112869737A CN 112869737 A CN112869737 A CN 112869737A CN 202110136278 A CN202110136278 A CN 202110136278A CN 112869737 A CN112869737 A CN 112869737A
Authority
CN
China
Prior art keywords
oxygen saturation
blood oxygen
signal
acquiring
dimension
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.)
Granted
Application number
CN202110136278.1A
Other languages
Chinese (zh)
Other versions
CN112869737B (en
Inventor
吴健
姜晓红
应豪超
曹燕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Industrial Technology Research Institute of ZJU
Original Assignee
Shandong Industrial Technology Research Institute of ZJU
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong Industrial Technology Research Institute of ZJU filed Critical Shandong Industrial Technology Research Institute of ZJU
Priority to CN202110136278.1A priority Critical patent/CN112869737B/en
Publication of CN112869737A publication Critical patent/CN112869737A/en
Application granted granted Critical
Publication of CN112869737B publication Critical patent/CN112869737B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
    • A61B5/1455—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue using optical sensors, e.g. spectral photometrical oximeters
    • A61B5/14551—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue using optical sensors, e.g. spectral photometrical oximeters for measuring blood gases
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161—Detection; Localisation; Normalisation
    • G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Pathology (AREA)
  • Signal Processing (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Medical Informatics (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Optics & Photonics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Human Computer Interaction (AREA)
  • Geometry (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physiology (AREA)
  • Psychiatry (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention belongs to the technical field of blood oxygen detection, and particularly relates to a non-contact human body blood oxygen saturation detection method. A non-contact human body blood oxygen saturation detection method comprises the following steps: s1, video acquisition; s2, acquiring an image area; s3, acquiring an original signal; s4, acquiring a separation signal; s5, acquiring variables; s6, acquiring result, and extracting AC based on step S5R、ACBAnd DCR、DCBAnd calculating a blood oxygen saturation parameter R, and obtaining the blood oxygen saturation SPO2 according to the Lambert-beer law. The invention provides a method for solving the problem of the published methodIn the method, a non-contact human blood oxygen saturation detection method is adopted based on the problem that the accuracy of a measurement result is not high due to the fact that the signal-to-noise ratio of an original signal extracted by a red channel and a blue channel is poor.

Description

Non-contact human body blood oxygen saturation detection method
Technical Field
The invention belongs to the technical field of blood oxygen detection, and particularly relates to a non-contact human body blood oxygen saturation detection method.
Background
The blood oxygen saturation (SPO 2) is the percentage of the hemoglobin volume in human blood combined by oxygen to the total hemoglobin volume, and is an important clinical health monitoring index as an important parameter of respiratory cycle, which can estimate the oxygenation of lung and hemoglobin oxygen carrying capacity.
Conventional methods for measuring blood oxygen saturation are classified into invasive methods and non-invasive methods. The invasive method is to collect blood from a human body and analyze the blood with a blood gas analyzer to calculate a value of blood oxygen saturation. This method is cumbersome and does not allow continuous measurement, the most important drawback being the pain or infection risk to the user. The noninvasive method mainly uses a finger-clip oximeter, and needs to attach the oximeter to a finger of a person, and although the noninvasive method can continuously measure the value, the noninvasive method is obviously not suitable for users with skin wounds or twitch patients, and also causes discomfort for general users. In recent years, imaging photoplethysmography (IPPG) has been developed, and has been a hot research direction in the non-contact field due to its feature of not requiring data acquisition directly through human skin images by contact with human skin. The study of the blood oxygen saturation measurement method is carried out by scholars at home and abroad based on the advantage of IPPG, the content is realized based on dual-wavelength visible light, the study is generally in an exploration stage, and the device is expensive and not portable enough, so that the device is not beneficial to daily monitoring.
With the rapid development of the scientific age, devices equipped with cameras, such as notebooks and smart phones, have become almost indispensable tools for life, and based on image data collected by these color cameras, it has proven feasible to measure the blood oxygen saturation level by the IPPG technique. Some of the published documents and patents mention that the red and blue channels of the color camera video are used as a dual-wavelength combination for measuring the blood oxygen saturation, but because the resolution of these low-end cameras is low and the sensitivity is weak, if the pixel values of the red and blue channels are simply used as the measurement parameters, the noise interference is large and the measurement result is not accurate enough.
Disclosure of Invention
The invention aims to solve the technical problem of providing a non-contact human blood oxygen saturation detection method which solves the problems of poor signal-to-noise ratio of original signals extracted based on red and blue channels and low accuracy of measurement results in the disclosed method. Therefore, the invention adopts the following technical scheme:
a non-contact human body blood oxygen saturation detection method comprises the following steps:
s1, acquiring a video, and acquiring human face video data;
s2, acquiring image areas, namely intercepting each frame of image of the video in the step S1, and selecting two cheek parts of the face from the video frame image as ROI (region of interest);
s3, obtaining original signals, extracting the average pixel value P of the red channel of the ROI in the step S2RAnd blue channel average pixel value PBAs the original pulse wave signal;
s4, obtaining the separation signals, and respectively carrying out the pulse wave red channel signals P obtained in the step S3 by adopting a single-channel independent component analysis algorithm based on dynamic embeddingRAnd blue channel signal PBCarrying out blind source separation to obtain a reconstructed separation signal IRAnd IB；
S5, acquiring variables, and respectively extracting the reconstructed signals I obtained in the step S4RAnd IBAC variable ACR、ACBAnd a direct current variable DCR、DCB；
S6, acquiring result, and extracting AC based on step S5R、ACBAnd DCR、DCBAnd calculating a blood oxygen saturation parameter R, and obtaining the blood oxygen saturation SPO2 according to the Lambert-beer law.
The method adopts a Dynamic Embedding (DE) -based single-channel Independent Component Analysis (ICA) algorithm (DE-ICA) to respectively carry out blind source separation on the acquired red and blue channel signals, effectively eliminates noise irrelevant to BVP, further carries out blood oxygen saturation estimation of a dual-wavelength method according to the Lambert-beer law and improves the detection accuracy.
On the basis of the technical scheme, the invention can also adopt the following further technical scheme:
the step S4 further includes the steps of:
s41, selecting embedding dimension m and time delay delta, and respectively adding P acquired in step S3RAnd PBAnd performing phase space reconstruction to form two groups of multidimensional vectors. The ICA algorithm requires that the number of observation signals is larger than or equal to the number of source signals, so that one-dimensional signals P with the length of N are firstly and respectively usedRAnd PBAnd performing phase space reconstruction to form two groups of multidimensional vectors.
S42, when m is larger, firstly using a rapid ICA algorithm to reduce the dimension of the obtained embedded matrix P to k, then using the rapid ICA algorithm to perform blind source separation on the matrix subjected to dimension reduction, and separating out an independent component Q;
s43, selecting independent component Q related to BVP from QjMultiplying by the column vector B of the corresponding mixing matrix BjObtaining independent sub-components Pj＝∑biQiI ∈ j, j is the set of independent components associated with the BVP;
s44, component PjPerforming inverse reconstruction to obtain final one-dimensional separation signal
Figure BDA0002926797860000031
Further, the step S42 further includes separating out an independent component Q ═ AP, [ Q ═ Q1,Q2,...,Qn]TA is an n × k dimension unmixing matrix, P is a multidimensional vector, n is the number of separated independent signals, and n is k;
estimating a mixing matrix B ═ pinv (a), B ═ B1,b2,...,bk]TPinv represents the pseudo-inverse.
Further, the method for selecting the embedding dimension m and the time delay Δ in step S41 includes:
the embedding dimension m and the time delay delta are chosen such that the observed signal P, respectivelyRAnd PBBecome new delay variableThe formula is as follows:
P(t)＝[p(t),p(t+Δ),...,p(t+m-1)Δ],t＝1,2,...,N-(m-1)Δ
and combining the signals according to the formula to form a multi-dimensional embedded delay matrix signal, wherein the formula is as follows:
Figure BDA0002926797860000032
the time delay delta is taken as 1, the embedding dimension m is set according to the sampling frequency of the observation signal and the minimum frequency of the source signal,
Figure BDA0002926797860000041
wherein fps is an observation signal PRAnd PBThe sampling frequency of (2).
The fps is the frame rate of the camera and is 30 frames/second; f. ofLAnd taking the minimum frequency of the separated source signal, wherein m is 30-60. Wherein, in order to ensure that the obtained embedded matrix P can extract enough implicit information, fLTypically taking the minimum frequency of the separated source signal. Since the signal frequency associated with BVP is typically 1-2HZ, m is typically 30-60.
Further, when the dimension of the obtained embedded matrix P is reduced by using the fast ICA algorithm in step S42, a dimension k after dimension reduction is selected based on the eigenvalue contribution cumulative quantity.
Further, the inverse process of DE-ICA in step S44 can be expressed as the following formula:
Figure BDA0002926797860000042
wherein
Figure BDA0002926797860000043
ceil denotes rounding up. According to the formula
Figure BDA0002926797860000044
The sequence of (A) is PjPerforming reconstruction to one-dimensional separation signal
Figure BDA0002926797860000045
Repeated elements during reconstruction are replaced by means.
Extracting a reconstructed signal I in the step S5RAnd IBAC variable ACR、ACBAnd a direct current variable DCR、DCBThe method comprises the following steps:
for the reconstructed signal IRAnd IBFiltering by adopting a band-pass filter of 0.6-3 HZ to extract a frequency signal S related to the human arterial blood flow volume pulse BVPRAnd SB；
By calculating SRAnd SBStandard deviation of (A) to obtain the AC variable ACRAnd ACB；
By calculating SRAnd SBObtaining the direct current variable DCR、DCB。
The method for calculating the blood oxygen saturation value in step S6 includes:
SPO2＝A+B*R
wherein the content of the first and second substances,
Figure BDA0002926797860000051
B. and B is obtained by performing least square linear fitting on the obtained R sequence and the oximeter reference value.
Compared with the prior art, the invention has the following beneficial effects:
the acquired red and blue channel signals are respectively subjected to blind source separation by adopting a DE-ICA algorithm, noise irrelevant to BVP is effectively eliminated, and then blood oxygen saturation estimation of a dual-wavelength method is carried out according to the Lambert-beer law, so that the detection accuracy is improved.
Drawings
FIG. 1 is a schematic flow chart of a non-contact human blood oxygen saturation detection method according to the present invention;
FIG. 2 is a schematic diagram of human face ROI detection according to a non-contact human blood oxygen saturation detection method of the present invention;
FIG. 3 is a waveform comparison diagram after noise filtering by the DE-ICA algorithm of the non-contact human blood oxygen saturation detection method of the present invention;
FIG. 4 is a diagram illustrating the results of measuring the blood oxygen saturation in the natural state by using the non-contact human blood oxygen saturation detection method of the present invention;
FIG. 5 is a schematic diagram of the results of measuring the blood oxygen saturation level in breath holding state by using the non-contact human blood oxygen saturation level detection method of the present invention.
Detailed Description
In order to further understand the present invention, the following will specifically describe the non-contact human blood oxygen saturation detection method provided by the present invention with reference to the specific embodiments, but the present invention is not limited thereto, and the non-essential modifications and adjustments made by those skilled in the art under the core guidance of the present invention still belong to the protection scope of the present invention.
As shown in fig. 1-5, a non-contact human blood oxygen saturation detection method includes the following steps:
and S1, acquiring human face video data.
Specifically, the intelligent mobile phone with the common camera is placed 50cm in front of the face, so that the face is kept right opposite to the camera and is completely in a shot picture. The resolution of the recorded video is 540 x 960, the frame rate is 30, the RGB color space is provided, the recorded video is 10s, the format is stored in an mp4 format, and the operating environment is win7+ Python 3.
And S2, intercepting each frame of image in the video, and selecting two cheek parts of the face from the video frame image as regions of interest (ROI).
Specifically, a human face detector Dlib library is adopted to identify the human face in the video image and locate 68 feature points of the face, and two cheek ROIs are framed based on part of the 68 feature points.
S3, extracting the average pixel value P of the red channel of the ROIRAnd blue channel average pixel value PBAs the original pulse wave signal.
Specifically, RGB channel values of two ROIs in each frame image are extracted, and red and blue channel pixel values of the two ROIs are averaged, thereby generating a two-channel signal P based on the ROIsRAnd PB。
S4, as shown in FIG. 3, respectively applying DE-ICA algorithm to pulse wave red channel signals PRAnd blue channel signal PBCarrying out blind source separation to obtain a reconstructed separation signal IRAnd IB；
Further, step S4 includes the steps of:
s41 and ICA algorithm require that the number of observed signals is larger than or equal to the number of source signals, so that one-dimensional signals P with the length of N are firstly respectively usedRAnd PBAnd performing phase space reconstruction to form two groups of multidimensional vectors.
In particular, the embedding dimension m and the time delay Δ are chosen such that the observed signal P is respectivelyRAnd PBBecoming a new delay variable, the formula is as follows:
P(t)＝[p(t),p(t+Δ),...,p(t+m-1)Δ],t＝1,2,...,N-(m-1)Δ
and combining the signals according to the formula to form a multi-dimensional embedded delay matrix signal, wherein the formula is as follows:
Figure BDA0002926797860000071
wherein, the time delay Δ is generally 1, and the embedding dimension m can be set according to the sampling frequency of the observation signal and the minimum frequency of the source signal, and the formula is as follows:
Figure BDA0002926797860000072
wherein fps is an observation signal PRAnd PBThe sampling frequency of (2), here the frame rate of the camera, is 30 frames/second. To ensure that the acquired embedding matrix P can extract enough implicit information, fLTypically taking the minimum frequency of the separated source signal. Since BVP-related signals are typically 1-2HZ in frequency, m is typically 1-2HZ30-60, 30 in this embodiment.
S42, because the embedding dimension m is large, the fast ica (FastICA) algorithm may be used to reduce the dimension of the obtained embedding matrix P to k (in this embodiment, the k value is selected based on the eigenvalue contribution rate cumulant being greater than or equal to 90%), and then the FastICA algorithm is used to perform blind source separation on the matrix after dimension reduction, so as to separate out an independent component Q ═ AP, Q ═ Q1,Q2,...,Qn]TWhere a is an n × k dimensional unmixing matrix, and n is the number of separated independent signals, where n ═ k. The mixing matrix B ═ pinv (A) and B ═ B are estimated simultaneously1,b2,...,bk]TPinv represents the pseudo-inverse.
S43, selecting independent component Q related to BVP from QjMultiplying by the column vector B of the corresponding mixing matrix BjObtaining independent sub-components Pj＝∑biQiI ∈ j, j is the set of independent components associated with the BVP.
S44, component PjPerforming inverse reconstruction to obtain final one-dimensional separation signal
Figure BDA0002926797860000073
Specifically, the inverse process of DE-ICA can represent the following formula,
Figure BDA0002926797860000081
wherein
Figure BDA0002926797860000082
ceil denotes rounding up. According to the formula
Figure BDA0002926797860000083
The sequence of (A) is PjPerforming reconstruction to one-dimensional separation signal
Figure BDA0002926797860000084
Repeated elements during reconstruction are replaced by means.
S45, for the S102-step obtained one-dimensional signal PRAnd PBRespectively processing the above steps to obtain a reconstructed one-dimensional separation signal IRAnd IB。
S5, respectively aligning the reconstructed signals IRAnd IBFiltering with a 0.6-3 HZ band-pass filter to extract a frequency signal S related to arterial Blood Volume Pulse (BVP)RAnd SBBy calculating SRAnd SBStandard deviation of (A) to obtain the AC variable ACRAnd ACB(ii) a By calculating SRAnd SBObtaining the direct current variable DCRAnd DCB。
S6, calculating according to the Lambert-beer law to obtain the blood oxygen saturation SPO2, wherein the formula is as follows:
SPO2＝A+B*R
wherein the content of the first and second substances,
Figure BDA0002926797860000085
A. b can be obtained by performing least square linear fitting on the obtained R sequence and an oximeter reference value, and then the blood oxygen saturation can be predicted based on the value.
The method of the present invention was tested as shown in fig. 4. In this example, 8 subjects, 4 females and 4 males were selected, and the experimental environment and procedure were as follows: under natural light, the room temperature is 23 ℃, a testee is enabled to sit at a position which is about 50cm away from the camera to keep a natural breathing state, facial videos of 10s are recorded as sample data, and the method provided by the invention is adopted to measure and calculate the blood oxygen saturation based on the sample data. Recording was performed by holding the subject with a conventional finger clip oximeter (standard for medical instruments) and recording the average of the oximetry values over 10s as a reference. The result shows that the measurement precision of the method provided by the invention reaches a better standard.
As shown in fig. 5, one of the subjects was selected for breath-hold test without loss of generality. The experimental environment was the same as above except that the subject was allowed to breathe naturally for 20s, and held for 30s from 21s while the subject's blood oxygen saturation value was recorded as a reference value by holding a common finger clip oximeter. The result also shows that the measurement precision of the method provided by the invention reaches a better standard.
It should be noted that the present disclosure is not limited to the foregoing embodiments and may be appropriately changed without departing from the spirit of the present disclosure; for example, different regions of interest are selected on the face in S101; for example, the band-pass filtering in S104 selects a suitable frequency range, etc.
While the invention has been shown and described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the appended claims.

Claims (8)

1. A non-contact human body blood oxygen saturation detection method is characterized by comprising the following steps:
s1, acquiring a video, and acquiring human face video data;
s2, acquiring image areas, namely intercepting each frame of image of the video in the step S1, and selecting two cheek parts of the face from the video frame image as ROI (region of interest);
s3, obtaining original signals, extracting the average pixel value P of the red channel of the ROI in the step S2RAnd blue channel average pixel value PBAs the original pulse wave signal;
s4, obtaining the separation signals, and respectively carrying out the pulse wave red channel signals P obtained in the step S3 by adopting a single-channel independent component analysis algorithm based on dynamic embeddingRAnd blue channel signal PBCarrying out blind source separation to obtain a reconstructed separation signal IRAnd IB；
S5, acquiring variables, and respectively extracting the reconstructed signals I obtained in the step S4RAnd IBAC variable ACR、ACBAnd a direct current variable DCR、DCB；
S6, acquiring result, and extracting AC based on step S5R、ACBAnd DCR、DCBCalculating a bleeding oxygen saturation parameter R according to Lambert-beer's law, resulting in blood oxygen saturation SPO 2.
2. The method for detecting blood oxygen saturation level of human body according to claim 1, wherein said step S4 further includes the steps of:
s41, selecting embedding dimension m and time delay delta, and respectively adding P acquired in step S3RAnd PBPerforming phase space reconstruction to form two groups of multidimensional vectors;
s42, when m is larger, firstly using a rapid ICA algorithm to reduce the dimension of the obtained embedded matrix P to k, then using the rapid ICA algorithm to perform blind source separation on the matrix subjected to dimension reduction, and separating out an independent component Q;
s43, selecting independent component Q related to BVP from QjMultiplying by the column vector B of the corresponding mixing matrix BjObtaining independent sub-components Pj＝∑biQiI ∈ j, j is the set of independent components associated with the BVP;
s44, component PjPerforming inverse reconstruction to obtain final one-dimensional separation signal
Figure FDA0002926797850000011
3. The method as claimed in claim 2, wherein the step S42 further comprises separating out the independent components Q ═ AP, Q ═ Q1,Q2,...,Qn]TA is an n × k dimension unmixing matrix, P is a multidimensional vector, n is the number of separated independent signals, and n is k;
estimating a mixing matrix B ═ pinv (a), B ═ B1,b2,...,bk]TPinv represents the pseudo-inverse.
4. The method of claim 2, wherein the step S41 of selecting the embedded dimension m and the time delay Δ comprises:
the time delay delta is taken as 1, the embedding dimension m is set according to the sampling frequency of the observation signal and the minimum frequency of the source signal,
Figure FDA0002926797850000021
wherein fps is an observation signal PRAnd PBThe sampling frequency of (2).
5. The method as claimed in claim 4, wherein fps is a frame rate of the camera, which is 30 frames/second; f. ofLAnd taking the minimum frequency of the separated source signal, wherein m is 30-60.
6. The method according to claim 2, wherein in step S42, when the fast ICA algorithm is used to perform dimension reduction on the obtained embedded matrix P, the dimension k after dimension reduction is selected based on the eigenvalue contribution cumulative quantity.
7. The method of claim 1, wherein the step S5 of extracting the reconstructed signal I is performed by a non-contact methodRAnd IBAC variable ACR、ACBAnd a direct current variable DCR、DCBThe method comprises the following steps:
for the reconstructed signal IRAnd IBFiltering by adopting a band-pass filter of 0.6-3 HZ to extract a frequency signal S related to the human arterial blood flow volume pulse BVPRAnd SB；
By calculating SRAnd SBStandard deviation of (A) to obtain the AC variable ACRAnd ACB；
By calculating SRAnd SBObtaining the direct current variable DCR、DCB。
8. The method as claimed in claim 1, wherein the calculating method of the blood oxygen saturation value in step S6 comprises:
SPO2＝A+B*R
wherein the content of the first and second substances,
Figure FDA0002926797850000031
A. and B is obtained by performing least square linear fitting on the obtained R sequence and the oximeter reference value.
CN202110136278.1A 2021-02-01 2021-02-01 Non-contact human body blood oxygen saturation detection method Active CN112869737B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110136278.1A CN112869737B (en) 2021-02-01 2021-02-01 Non-contact human body blood oxygen saturation detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110136278.1A CN112869737B (en) 2021-02-01 2021-02-01 Non-contact human body blood oxygen saturation detection method

Publications (2)

Publication Number Publication Date
CN112869737A true CN112869737A (en) 2021-06-01
CN112869737B CN112869737B (en) 2022-10-28

Family

ID=76052262

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110136278.1A Active CN112869737B (en) 2021-02-01 2021-02-01 Non-contact human body blood oxygen saturation detection method

Country Status (1)

Country Link
CN (1) CN112869737B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114140859A (en) * 2021-12-08 2022-03-04 山东大学 Non-contact oxyhemoglobin saturation measuring method and system
CN115205270A (en) * 2022-07-25 2022-10-18 哈尔滨工业大学 Non-contact type blood oxygen saturation detection method and system based on image processing

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030088164A1 (en) * 2001-11-02 2003-05-08 Nellcor Puritan Bennett Inc. Blind source separation of pulse oximetry signals
CN101940475A (en) * 2010-09-03 2011-01-12 深圳市纽泰克电子有限公司 Method for improving detection accuracy of blood oxygen saturation
US20130072771A1 (en) * 2011-09-21 2013-03-21 Pixart Imaging Inc. Optical finger mouse, electronic device and physiological characteristics detection device
CN107358220A (en) * 2017-07-31 2017-11-17 江西中医药大学 A kind of human heart rate and the contactless measurement of breathing
CN107456219A (en) * 2017-09-07 2017-12-12 成都云卫康医疗科技有限公司 One kind is based on Pearson correlation coefficient Dynamic Heart Rate and blood oxygen measuring method
CN107773223A (en) * 2016-08-24 2018-03-09 上海佑壳尔科技有限公司 A kind of online finger instrument for obtaining pulse wave and acquisition pulse wave peak shape parameter method
CN110236515A (en) * 2019-07-19 2019-09-17 合肥工业大学 A kind of contactless heart rate detection method based on near-infrared video
CN110755090A (en) * 2019-11-15 2020-02-07 合肥师范学院 Human body blood oxygen saturation measuring method based on network camera
CN111444797A (en) * 2020-03-16 2020-07-24 五邑大学 Non-contact heart rate detection method
CN111714105A (en) * 2020-07-24 2020-09-29 长春理工大学 Human vital sign perception system based on IPPG
CN112116922A (en) * 2020-09-17 2020-12-22 集美大学 Noise blind source signal separation method, terminal equipment and storage medium
CN112200099A (en) * 2020-10-14 2021-01-08 浙江大学山东工业技术研究院 Video-based dynamic heart rate detection method
CN112237421A (en) * 2020-09-23 2021-01-19 浙江大学山东工业技术研究院 Video-based dynamic heart rate variability analysis model

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030088164A1 (en) * 2001-11-02 2003-05-08 Nellcor Puritan Bennett Inc. Blind source separation of pulse oximetry signals
CN101940475A (en) * 2010-09-03 2011-01-12 深圳市纽泰克电子有限公司 Method for improving detection accuracy of blood oxygen saturation
US20130072771A1 (en) * 2011-09-21 2013-03-21 Pixart Imaging Inc. Optical finger mouse, electronic device and physiological characteristics detection device
CN107773223A (en) * 2016-08-24 2018-03-09 上海佑壳尔科技有限公司 A kind of online finger instrument for obtaining pulse wave and acquisition pulse wave peak shape parameter method
CN107358220A (en) * 2017-07-31 2017-11-17 江西中医药大学 A kind of human heart rate and the contactless measurement of breathing
CN107456219A (en) * 2017-09-07 2017-12-12 成都云卫康医疗科技有限公司 One kind is based on Pearson correlation coefficient Dynamic Heart Rate and blood oxygen measuring method
CN110236515A (en) * 2019-07-19 2019-09-17 合肥工业大学 A kind of contactless heart rate detection method based on near-infrared video
CN110755090A (en) * 2019-11-15 2020-02-07 合肥师范学院 Human body blood oxygen saturation measuring method based on network camera
CN111444797A (en) * 2020-03-16 2020-07-24 五邑大学 Non-contact heart rate detection method
CN111714105A (en) * 2020-07-24 2020-09-29 长春理工大学 Human vital sign perception system based on IPPG
CN112116922A (en) * 2020-09-17 2020-12-22 集美大学 Noise blind source signal separation method, terminal equipment and storage medium
CN112237421A (en) * 2020-09-23 2021-01-19 浙江大学山东工业技术研究院 Video-based dynamic heart rate variability analysis model
CN112200099A (en) * 2020-10-14 2021-01-08 浙江大学山东工业技术研究院 Video-based dynamic heart rate detection method

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114140859A (en) * 2021-12-08 2022-03-04 山东大学 Non-contact oxyhemoglobin saturation measuring method and system
CN115205270A (en) * 2022-07-25 2022-10-18 哈尔滨工业大学 Non-contact type blood oxygen saturation detection method and system based on image processing
CN115205270B (en) * 2022-07-25 2023-10-24 哈尔滨工业大学 Non-contact blood oxygen saturation detection method and system based on image processing

Also Published As

Publication number Publication date
CN112869737B (en) 2022-10-28

Similar Documents

Publication Publication Date Title
Wang et al. A comparative survey of methods for remote heart rate detection from frontal face videos
Tasli et al. Remote PPG based vital sign measurement using adaptive facial regions
KR101629901B1 (en) Method and Device for measuring PPG signal by using mobile device
CN112869737B (en) Non-contact human body blood oxygen saturation detection method
Gudi et al. Efficient real-time camera based estimation of heart rate and its variability
Feng et al. Motion artifacts suppression for remote imaging photoplethysmography
CN112233813A (en) Non-contact non-invasive heart rate and respiration measurement method and system based on PPG
CN111387959A (en) Non-contact physiological parameter detection method based on IPPG
CN112294282A (en) Self-calibration method of emotion detection device based on RPPG
Li et al. An improvement for video-based heart rate variability measurement
CN112200099A (en) Video-based dynamic heart rate detection method
CN113591769B (en) Non-contact heart rate detection method based on photoplethysmography
Nikolic-Popovic et al. Impact of motion artifacts on video-based non-intrusive heart rate measurement
Lamba et al. Contactless heart rate estimation from face videos
WO2014030091A1 (en) Method and apparatus for measuring physiological parameters of an object
Jhuang et al. A patch-sized wearable ECG/respiration recording platform with DSP capability
Karmuse et al. A robust rPPG approach for continuous heart rate measurement based on face
Mehta et al. CPulse: Heart rate estimation from RGB videos under realistic conditions
CN107280673A (en) A kind of infrared imaging breath signal detection method based on key-frame extraction technique
CN113693573B (en) Video-based non-contact multi-physiological-parameter monitoring system and method
CN104688199A (en) Non-contact type pulse measurement method based on skin pigment concentration difference
Kuang et al. Shuffle-rPPGNet: Efficient network with global context for remote heart rate variability measurement
Yue et al. A new approach for noncontact imaging photoplethysmography using chrominance features and low-rank in the IoT operating room
Maji et al. Contactless heart rate monitoring using A standard RGB camera
AV et al. Non-contact heart rate monitoring using machine learning

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
GR01 Patent grant
GR01 Patent grant