CN105046209B - A kind of contactless method for measuring heart rate based on canonical correlation analysis - Google Patents
A kind of contactless method for measuring heart rate based on canonical correlation analysis Download PDFInfo
- Publication number
- CN105046209B CN105046209B CN201510373448.2A CN201510373448A CN105046209B CN 105046209 B CN105046209 B CN 105046209B CN 201510373448 A CN201510373448 A CN 201510373448A CN 105046209 B CN105046209 B CN 105046209B
- Authority
- CN
- China
- Prior art keywords
- heart rate
- channel
- value
- signal
- image
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- 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
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
A kind of contactless method for measuring heart rate based on canonical correlation analysis, steps are as follows: 1) acquiring face video, carry out ROI framing extraction;2) ROI region that every frame is extracted, carry out three primary colours separation, generate red, green, blue triple channel image, and mean value is taken to all pixels value of each channel image of the frame image, signal value of the mean value as the frame image in the channel, to generate the original signal sequence in three channels of measured zone image;3) original signal sequence in three channels is carried out respectively linearizing and standardizing;4) according to standardization postamble sequence and template Y, seek vector W respectivelyxWith WySo that x, y are in vector WxAnd WyOn projection X=xTWx, Y=yTWyBetween correlation ρ it is maximum;5) using the signal after the standardization in the maximum channel of correlation coefficient ρ as heart rate signal.Present invention eliminates the processes of complicated independent principal component analysis and signal processing, simplify the process of measurement, therefore shorten the used time of measurement.
Description
Technical field
The present invention relates to heart rate measurement field, especially a kind of contactless heart rate measurement side based on canonical correlation analysis
Method.
Background technique
Heart rate can be obtained from electrocardiosignal and pulse.The acquisition of electrocardiosignal needs cardiac electrical by special measurement
Instrument obtains;Pulse frequency and heart rate are consistent in human normal, can estimate heart rate using pulse frequency, in traditional
Doctor feels the pulse, sphygmomanometer, finger tip pulse transducer and some photoelectric intelligent pulses measure instruments.But the common ground of these methods
It is that the measured and instrument or doctor is required to have a degree of physical contact, these contacts are a degree of to measurement band
Inconvenience also easily causes the measured discomfort.
Summary of the invention
It is a primary object of the present invention to overcome drawbacks described above in the prior art, proposes a kind of simplified step, shortens and use
When the contactless method for measuring heart rate based on canonical correlation analysis.
The present invention adopts the following technical scheme:
A kind of contactless method for measuring heart rate based on canonical correlation analysis, it is characterised in that: pre-establish heart rate letter
Number template Y, steps are as follows for remaining:
1) face video is acquired, then selectes the face specific region in face video as measured zone, then carry out ROI
Framing is extracted;
2) ROI region extracted to every frame carries out three primary colours separation, generates red, green, blue triple channel image, and to this
The all pixels value of each channel image of frame image takes mean value, signal value of the mean value as the frame image in the channel, from
And generate the original signal sequence X in three channels of measured zone imageR(t)、XG(t) and XB(t);
3) by the original signal sequence X in three channelsR(t), XG(t) and XB(t) carry out linearizing respectively, then respectively into
Row standardization, it is ensured that the amplitude of the original signal sequence in each channel is in specific sections, and after being standardizedWithWherein:The original signal value after linearisation is gone in X (t) expression,Indicate former
The average value of beginning signal sequence, S indicate the standard deviation square value of original signal sequence;
4) according to, with template Y, seeking vector W respectively after standardizationxWith WySo that x, y are in vector Wx
And WyOn projection X=xTWx, Y=yTWyBetween correlation ρ it is maximum,
E [] indicates expectation;
5) using the signal after the standardization in the maximum channel of correlation coefficient ρ as heart rate signal.
Preferably, the signal templates are as follows: Y=α1sin(2π*50*t/60)+α2Cos (2 π * 50*t/60), t takes and heart rate
Signal isometric time, α1, α2For the real number that cannot be simultaneously 0.
Preferably, in step 1), the face specific region refers to eyebrow top and lip lower edge for height, cheek
It is the rectangle of width on the outside of both sides.
Preferably, in step 3), described goes linearisation to realize using the detrend function that Mat lab is carried.
By the above-mentioned description of this invention it is found that compared with prior art, the invention has the following beneficial effects:
1, the present invention is based on the contactless method for measuring heart rate of canonical correlation analysis, pass through frequency point compared to traditional
The method of analysis eliminates the process of complicated independent principal component analysis and signal processing, simplifies the process of measurement, therefore shorten
Used time of measurement.
2, in the accuracy of measurement, the method for canonical correlation analysis mean error in the data surveyed.Pass through
The analysis of Bland-Altman ratio and difference, most numerical value is all in the preferable section of consistency, it can be seen that measurement
There is good consistency between value and true value.
Detailed description of the invention
Fig. 1 is flow chart of the invention;
Fig. 2 is the schematic diagram of present invention acquisition face video;
Fig. 3 is the schematic diagram of measured zone of the present invention;
Fig. 4 is three primary colours seperated schematic diagram;
Fig. 5 is each channel heart rate value of subject and practical heart rate tested using the method for the present invention;
Fig. 6 is green channel figure compared with practical heart rate.
Specific embodiment
Below by way of specific embodiment, the invention will be further described.
A kind of contactless method for measuring heart rate based on canonical correlation analysis of the invention, principle is: heartbeat
Caused dermovascular volume variation, blood is different to the absorption of the light beam of different-waveband, causes the variation of reflected light,
Reflected light is able to reflect time change and its period of cardiovascular activity cardiac beating, i.e. heart rate information.It is received with camera
Reflected light forms color video frequency image, brightness change of the every frame image in collected video in three Color Channel of red, green, blue
Raw digital signal is formed, canonical correlation analysis is directly carried out to raw digital signal so that it is determined that heart rate value.Applied to mobile phone
The heart rate value of human body can be more quickly obtained on equal portable devices.
Since certain periodicity is presented in heart rate signal, for 50 beats/min of heart rate, here 50 beats/min of heart rate
Signal template set up by taking sinusoidal signal as an example, it is contemplated that the variation of phase considers to be superimposed a cosine signal template here public
Formula is as follows:
Y=(sin (2 π * 50*t/60), cos (2 π * 50*t/60))
Template will carry out in the following manner linear combination:
α1sin(2π*50*t/60)+α2cos(2π*50*t/60)
Wherein t is taken as the time isometric with heart rate signal, α1, α2It is the real number that cannot be simultaneously 0.
Referring to Fig.1, remaining steps are as follows:
1) face video is acquired, face is shot using 30 frames/second video camera (or mobile phone), makes face as far as possible
Often have in camera lens (referring to Fig. 2), then selectes the face specific region in face video as measured zone (referring to figure
3), then ROI framing extraction is carried out.Due to, before research shows that four different parts (forehead (ROI to same subject
I), the zonule (ROI II) in ROI I, local hair zones (ROI III) and entire head zone (ROI IV)) it is right
Than analysis, discovery ROI I, ROI II and ROI IV difference are little, illustrate that requirement of the heart rate detection to regional choice is not very
Strictly, but to include entire human face region measurement effect preferably and have very high identification.Therefore non-contact measurement is selected
Position based on face is as measured zone.ROI region is chosen for eyebrow top and lip lower edge as height, cheek by us
It is the rectangle (can suitably include non-face area) of width on the outside of both sides.
2) three primary colours separation is carried out to the ROI region that every frame extracts referring to Fig. 4, generates red, green, blue triple channel figure
Picture, and mean value is taken to all pixels value of each channel image of the frame image, the mean value is as the frame image in the channel
Signal value, to generate the original signal sequence X in three channels of measured zone imageR(t)、XG(t) and XB(t)。
3) by the original signal sequence X in three channelsR(t), XG(t) and XB(t) it carries out linearizing respectively, goes to linearize
The detrend function carried using Matlab.Since heart rate signal is in undulating, the original signal after going linearisation need to be marked
Standardization (normalization), it is ensured that the amplitude of the original signal sequence in each channel is in specific sections, and after being standardizedWithWherein:The original signal sequence value after linearisation is gone in X (t) expression,Table
Show that the average value of original signal sequence, S indicate the standard deviation square value of original signal sequence.
4) canonical correlation analysis, it is contemplated that heart rate range under normal circumstances, we walk provided with 50Hz-120Hz here
A length of 1Hz totally 71 template signals, the 1st template correspond to 50Hz, and so on the 71st template correspond to 120Hz.With
The original signal sequence in each channel measured carries out canonical correlation calculating with each template signal respectively, obtains degree of correlation most
High stencil value.It is specific: according to, with template Y, wherein template is Y=(sin (2 π * after standardization
50*t/60),cos(2π*50*t/60));Seek vector W respectivelyxWith WySo that x, y are in vector WxAnd WyOn projection X=
xTWx, Y=yTWyBetween correlation ρ it is maximum,E [] is indicated
It is expected that;
5) using the signal after the standardization in the maximum channel of correlation coefficient ρ as heart rate signal.
Fig. 5 is each channel original signal sequence measured using the method for the present invention and practical heart rate value, from each by
In the data that examination person measures, will all the correlation between three channels of red, green, blue and each template be obtained, takes related coefficient maximum
Value as measuring heart rate signal.It, can be with through each channel heart rate measurements (original signal sequence) compared with true value
Find out that the value in green channel will be significantly better than other channels.Straight line in Fig. 6 is the solid line that slope is 1, it can be found that this method
The heart rate value measured on green channel is all distributed in the two sides close to straight line substantially, and it is preferable related to show that this two groups of data have
Property.Finally using the normalized signal in green channel as the channel of heart rate measurement.
The above is only a specific embodiment of the present invention, but the design concept of the present invention is not limited to this, all to utilize this
Design makes a non-material change to the present invention, and should all belong to behavior that violates the scope of protection of the present invention.
Claims (3)
1. a kind of contactless method for measuring heart rate based on canonical correlation analysis, it is characterised in that: pre-establish heart rate signal
Template Y, Y=α1sin(2π*50*t/60)+α2Cos (2 π * 50*t/60), t takes the time isometric with heart rate signal, α1, α2For not
Can simultaneously be 0 real number, steps are as follows for remaining:
1) face video is acquired, then selectes the face specific region in face video as measured zone, then carry out ROI framing
It extracts;
2) ROI region extracted to every frame carries out three primary colours separation, generates red, green, blue triple channel image, and to the frame figure
The all pixels value of each channel image of picture takes mean value, signal value of the mean value as the frame image in the channel, thus raw
At the original signal sequence X in three channels of measured zone imageR(t)、XG(t) and XB(t);
3) by the original signal sequence X in three channelsR(t), XG(t) and XB(t) it carries out linearizing respectively, then is marked respectively
Standardization, it is ensured that the amplitude of the original signal sequence in each channel is in specific sections, and after being standardizedWithWherein:The original signal value after linearisation is gone in X (t) expression,Indicate original
The average value of signal sequence, S indicate the standard deviation square value of original signal sequence;
4) after according to standardizationWith template Y, seek vector W respectivelyxWith WySo that x, y are in vector WxAnd Wy
On projection X=xTWx, Y=yTWyBetween correlation ρ it is maximum,
E [] indicates expectation;
5) using the signal after the standardization in the maximum channel of correlation coefficient ρ as heart rate signal.
2. a kind of contactless method for measuring heart rate based on canonical correlation analysis as described in claim 1, it is characterised in that:
In step 1), it is width on the outside of cheek both sides that the face specific region, which refers to eyebrow top and lip lower edge for height,
Rectangle.
3. a kind of contactless method for measuring heart rate based on canonical correlation analysis as described in claim 1, it is characterised in that:
In step 3), described goes linearisation to realize using the detrend function that Matlab is carried.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510373448.2A CN105046209B (en) | 2015-06-30 | 2015-06-30 | A kind of contactless method for measuring heart rate based on canonical correlation analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510373448.2A CN105046209B (en) | 2015-06-30 | 2015-06-30 | A kind of contactless method for measuring heart rate based on canonical correlation analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105046209A CN105046209A (en) | 2015-11-11 |
CN105046209B true CN105046209B (en) | 2019-01-25 |
Family
ID=54452738
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510373448.2A Active CN105046209B (en) | 2015-06-30 | 2015-06-30 | A kind of contactless method for measuring heart rate based on canonical correlation analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105046209B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105615864A (en) * | 2016-03-31 | 2016-06-01 | 四川东鼎里智信息技术有限责任公司 | Information collection method for mobile terminal |
CN105962915B (en) * | 2016-06-02 | 2019-04-05 | 安徽大学 | Contactless humanbody respiratory rate and heart rate method for synchronously measuring and system |
CN108937905B (en) * | 2018-08-06 | 2021-05-28 | 合肥工业大学 | Non-contact heart rate detection method based on signal fitting |
CN110269600B (en) * | 2019-08-06 | 2021-12-21 | 合肥工业大学 | Non-contact video heart rate detection method based on multivariate empirical mode decomposition and combined blind source separation |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102499664A (en) * | 2011-10-24 | 2012-06-20 | 西双版纳大渡云海生物科技发展有限公司 | Video-image-based method and system for detecting non-contact vital sign |
CN103702014A (en) * | 2013-12-31 | 2014-04-02 | 中国科学院深圳先进技术研究院 | Non-contact physiological parameter detection method, system and device |
CN104639799A (en) * | 2013-11-14 | 2015-05-20 | 财团法人工业技术研究院 | Image type heart rate activity detection device and method thereof |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110251493A1 (en) * | 2010-03-22 | 2011-10-13 | Massachusetts Institute Of Technology | Method and system for measurement of physiological parameters |
-
2015
- 2015-06-30 CN CN201510373448.2A patent/CN105046209B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102499664A (en) * | 2011-10-24 | 2012-06-20 | 西双版纳大渡云海生物科技发展有限公司 | Video-image-based method and system for detecting non-contact vital sign |
CN104639799A (en) * | 2013-11-14 | 2015-05-20 | 财团法人工业技术研究院 | Image type heart rate activity detection device and method thereof |
CN103702014A (en) * | 2013-12-31 | 2014-04-02 | 中国科学院深圳先进技术研究院 | Non-contact physiological parameter detection method, system and device |
Non-Patent Citations (2)
Title |
---|
A CANONICAL CORRELATION APPROACH TO HEART BEAT DETECTION IN TEXTILE ECG MEASUREMENTS;L. Rattfalt 等;《IET 3rd International Conference On Advances in Medical,Signal and Information Processing. MEDSIP 2006》;20060719;第1页摘要,第2-3页Methodlogy-Result |
Canonical correlation analysis:An overview with application to learning methods;David R. Hardoon;《Neural Computation》;20041201;第16卷(第12期);2639-2664 |
Also Published As
Publication number | Publication date |
---|---|
CN105046209A (en) | 2015-11-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Feng et al. | Motion-resistant remote imaging photoplethysmography based on the optical properties of skin | |
Lam et al. | Robust heart rate measurement from video using select random patches | |
CN110269600B (en) | Non-contact video heart rate detection method based on multivariate empirical mode decomposition and combined blind source separation | |
Shao et al. | Noncontact monitoring breathing pattern, exhalation flow rate and pulse transit time | |
JP6371837B2 (en) | Devices and methods for obtaining vital signs of subjects | |
CN103908236B (en) | A kind of automatic blood pressure measurement system | |
US20170112382A1 (en) | Pulse-wave detection method, pulse-wave detection device, and computer-readable recording medium | |
CN105147274B (en) | A kind of method that heart rate is extracted in the face video signal from visible spectrum | |
US20200121262A1 (en) | Device, system and method for measuring and processing physiological signals of a subject | |
CN105046209B (en) | A kind of contactless method for measuring heart rate based on canonical correlation analysis | |
JP6521845B2 (en) | Device and method for measuring periodic fluctuation linked to heart beat | |
Feng et al. | Motion artifacts suppression for remote imaging photoplethysmography | |
CN104138254A (en) | Non-contact type automatic heart rate measurement system and measurement method | |
JP6098304B2 (en) | Pulse wave detection device, pulse wave detection method, and pulse wave detection program | |
CN109313934B (en) | CPR assist device and method for determining depth of chest compressions in a patient | |
CN105989357A (en) | Human face video processing-based heart rate detection method | |
TW201315438A (en) | Method of contact-free heart rate estimation and system thereof | |
CN104173051A (en) | Automatic noncontact respiration assessing system and assessing method | |
CN203290905U (en) | Non-contact automatic respiration measuring system | |
KR102358325B1 (en) | Method and apparatus for measuring robust pulse rate and respiratory rate using face images | |
JP2020178964A (en) | Biological information detection device, biological information detection method and biological information detection program | |
US11701015B2 (en) | Computer-implemented method and system for direct photoplethysmography (PPG) with multiple sensors | |
Cui et al. | Non-contact time varying heart rate monitoring in exercise by video camera | |
Jensen et al. | Camera-based heart rate monitoring | |
Karlen et al. | Respiratory rate assessment from photoplethysmographic imaging |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |