WO2022077887A1 - 一种基于视频数据进行血压预测的系统 - Google Patents

一种基于视频数据进行血压预测的系统 Download PDF

Info

Publication number
WO2022077887A1
WO2022077887A1 PCT/CN2021/088017 CN2021088017W WO2022077887A1 WO 2022077887 A1 WO2022077887 A1 WO 2022077887A1 CN 2021088017 W CN2021088017 W CN 2021088017W WO 2022077887 A1 WO2022077887 A1 WO 2022077887A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
information
generate
blood pressure
perform
Prior art date
Application number
PCT/CN2021/088017
Other languages
English (en)
French (fr)
Inventor
曹君
吴泽剑
王思瀚
张碧莹
孙洪岱
李喆
章诚忠
李瑞莱
臧凯丰
鲁瑞广
王月雷
班亮
赵晓乐
齐昕
Original Assignee
乐普（北京）医疗器械股份有限公司
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 乐普（北京）医疗器械股份有限公司 filed Critical 乐普（北京）医疗器械股份有限公司
Priority to US18/248,713 priority Critical patent/US20230389808A1/en
Priority to EP21878941.0A priority patent/EP4226855A1/en
Publication of WO2022077887A1 publication Critical patent/WO2022077887A1/zh

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
    • A61B5/02—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
    • A61B5/021—Measuring pressure in heart or blood vessels
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
    • 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
    • A61B5/021—Measuring pressure in heart or blood vessels
    • A61B5/02108—Measuring pressure in heart or blood vessels from analysis of pulse wave characteristics
    • A61B5/02116—Measuring pressure in heart or blood vessels from analysis of pulse wave characteristics of pulse wave amplitude
    • 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
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A61B5/02416—Detecting, measuring or recording pulse rate or heart rate using photoplethysmograph signals, e.g. generated by infrared radiation
    • 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
    • A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
    • 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
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
    • 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
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/46—Interconnection of networks
    • H04L12/4633—Interconnection of networks using encapsulation techniques, e.g. tunneling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/12—Applying verification of the received information
    • H04L63/123—Applying verification of the received information received data contents, e.g. message integrity
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/56—Provisioning of proxy services
    • H04L67/565—Conversion or adaptation of application format or content
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/32—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
    • H04L9/321—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving a third party or a trusted authority
    • H04L9/3213—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving a third party or a trusted authority using tickets or tokens, e.g. Kerberos
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L2209/00—Additional information or applications relating to cryptographic mechanisms or cryptographic arrangements for secret or secure communication H04L9/00
    • H04L2209/88—Medical equipments

Definitions

  • the invention relates to the technical field of signal processing, in particular to a system for blood pressure prediction based on video data.
  • Photoplethysmography is a non-invasive detection method that detects blood volume changes in living tissue by means of photoelectric means.
  • the heart beat will make the blood flow per unit area in the blood vessel to form a periodic change, and the corresponding blood volume will also change accordingly, so that the photoplethysmographic signal reflecting the amount of light absorbed by the blood will also change periodically.
  • the periodic change of the meter signal is closely related to the heart beat and blood pressure change.
  • the heart rate data of the heart can be obtained by performing heartbeat interval analysis on the conventional photoplethysmographic signal; the blood pressure correlation analysis can be performed on the conventional photoplethysmographic signal using a well-trained artificial intelligence blood pressure prediction model, and the diastolic and systolic blood pressure can be obtained.
  • the remote photoplethysmography signal is a periodic signal produced by the absorption or reflection of light by the skin.
  • the heart rate calculation and blood pressure analysis based on the remote photoplethysmographic signal are the same as those based on the photoplethysmographic signal.
  • the test object needs to wear a customized acquisition device (for example, acquisition equipment such as finger clips, ear clips, etc.), which is not convenient for the test object to detect itself immediately, nor to the test object. Perform continuous monitoring.
  • acquisition equipment for example, acquisition equipment such as finger clips, ear clips, etc.
  • the purpose of the present invention is to aim at the defects of the prior art.
  • the present invention provides a system for blood pressure prediction based on video data, including a first device and a cloud server; using this system, it is possible to pass the system without wearing any specific collection device.
  • the first device enables real-time detection and continuous monitoring, which can not only reduce the difficulty of real-time blood pressure detection and continuous monitoring of the test object, but also enrich the application scenarios of the photoplethysmography method in the monitoring field.
  • the present invention provides a system for blood pressure prediction based on video data, including: a first device and a cloud server;
  • the first device includes a first main control module, a first camera, a lighting module, a display screen and a first communication module;
  • the first main control module is used to call the first camera and the lighting module to perform a first duration of shooting processing on the epidermis area of the test object, thereby generating first video data;
  • the display screen is used for receiving the first video data sent by the main control module, and performing playback processing
  • the first main control module is further configured to perform light source channel data extraction processing on the first video data according to the light source information to generate first channel data; and then perform a remote photoplethysmography signal on the first channel data Data conversion processing to generate first signal data; and then sending the first signal data to the display screen for signal waveform display processing according to the display time length;
  • the first main control module is further configured to store the first signal data, the first device token information, the first device type information, the first age information, the first gender information, and the first height information of the test subject. and the first weight information, encapsulated into a first data packet according to a first protocol, and using the first communication module to send the first data packet to the cloud server;
  • the cloud server includes a second main control module, a legality verification module, a parameter verification module, a data preprocessing module, an artificial intelligence blood pressure prediction module and a second communication module;
  • the second main control module is configured to perform data parsing processing on the first data packet according to the first protocol to obtain second signal data, second device token information, second device type information, and second age information , second gender information, second height information and second weight information;
  • the legality verification module is used to perform legality verification processing on the token information of the second device according to the legal token list;
  • the parameter verification module is configured to, when the legitimacy verification process is successful, verify the second signal data, the second device type information, the second age information, the second gender information, the second height information and the second weight information, and perform parameter verification processing;
  • the second main control module is further configured to perform heart rate calculation processing according to the second signal data to generate heart rate data when the parameter integrity verification processing is successful;
  • the data preprocessing module is used for, according to the prediction model identification information, the second signal data, the second age information, the second gender information, the second height information and the second weight information, Prepare and process the input data of the blood pressure prediction model, and generate the model input data;
  • the artificial intelligence blood pressure prediction module is configured to input data to the model according to the prediction model identification information, perform blood pressure prediction calculation processing, and generate diastolic blood pressure data and systolic blood pressure data;
  • the second main control module is also used to set the status code data as normal status code information, and then form return data according to the heart rate data, the diastolic blood pressure data and the systolic blood pressure data, and then combine the returned data with the all data.
  • the status code data is encapsulated into a second data packet according to the first protocol, and the second data packet is sent to the first device by the second communication module;
  • the first main control module is further configured to perform data parsing processing on the second data packet according to the first protocol to obtain the returned data and the status code data; when the status code data is the normal When the status code information is received, obtain the heart rate data, the diastolic blood pressure data and the systolic blood pressure data from the returned data; then send the heart rate data, the diastolic blood pressure data and the systolic blood pressure data to the Heart rate and blood pressure data display processing is performed on the above-mentioned display screen.
  • the first main control module is specifically configured to call the lighting module to illuminate the skin area of the test object, and after the lens of the first camera covers the skin area, use the first camera to illuminate the skin area.
  • the epidermal region is subjected to a photographing process of the first duration, thereby generating the first video data.
  • the first main control module is specifically configured to perform red light channel data extraction processing on the first video data when the light source information is red light to generate the first channel data; when the light source information is green light When the light source is light, perform green channel data extraction processing on the first video data to generate the first channel data; when the light source information is red and green light, perform red channel data extraction on the first video data Process to generate first red channel data, perform green channel data extraction processing on the first video data to generate first green channel data, and combine the first red channel data and the first green channel data packaged into the first channel data.
  • the first main control module is specifically configured to perform frame image extraction processing on the first video data when the light source information is the red light to obtain a plurality of first frame image data; In the image data, count the number of the first red pixel points whose pixel value meets the red light pixel threshold range to generate a first total number, and perform a pixel value sum calculation for all the first red pixel points to generate a first pixel value. sum, and then use the ratio of the sum of the first pixel values to the first total number as the first frame of red light channel data corresponding to each first frame of image data; then for all the first frames
  • the red light channel data is arranged in chronological order to generate the first channel data;
  • the first main control module is specifically configured to perform frame image extraction processing on the first video data when the light source information is the green light to obtain a plurality of second frame image data; In the image data, count the number of first green pixel points whose pixel value meets the green pixel threshold range to generate a second total number, and perform pixel value sum calculation on all the first green pixel points to generate a second pixel value sum, and then take the ratio of the sum of the second pixel values to the second total as the first frame of green light channel data corresponding to each second frame of image data; then for all the first frames Green light channel data, arranged in chronological order to generate the first channel data;
  • the first main control module is specifically configured to perform video frame image extraction processing on the first video data when the light source information is the red and green light to obtain a plurality of third frame image data; In the three frames of image data, count the number of second red pixels whose pixel values satisfy the red light pixel threshold range to generate a third total number, and calculate the sum of pixel values for all the second red pixels to generate The sum of the third pixel values, and then the ratio of the sum of the third pixel values to the third total is taken as the second frame of red light channel data corresponding to each of the third frames of image data, and then for all The second frame of red light channel data is arranged in chronological order to generate the first red light channel data; and in each of the third frames of image data, the pixel values that satisfy the green light pixel threshold range are determined.
  • the number of the second green pixel points is counted to generate a fourth total number, and the pixel value sum calculation is performed on all the second green pixel points to generate a fourth pixel value sum, and then the fourth pixel value sum is added to the
  • the ratio of the fourth total number is used as the second frame of green light channel data corresponding to each third frame of image data, and then all the second frames of green light channel data are arranged in chronological order to generate the the first green light channel data; and then perform multi-channel data encapsulation processing on the first red light channel data and the first green light channel data to generate the first channel data.
  • the first main control module is specifically configured to perform remote photoplethysmography signal bandpass filtering processing on the first channel data when the light source information is the red light, generate first red light filtering data, and then generate the first red light filtering data. performing remote photoplethysmography signal noise reduction processing on the first red light filtering data to generate first red light signal data;
  • the first main control module is specifically configured to perform remote photoplethysmography signal bandpass filtering processing on the first channel data when the light source information is the green light, generate first green light filtering data, and then generate the first green light filtering data. performing remote photoplethysmography signal noise reduction processing on the first green light filtering data to generate first green light signal data;
  • the first main control module is specifically configured to perform red light channel data extraction processing on the first channel data when the light source information is the red and green light to generate second red light channel data;
  • the first channel data is subjected to green light channel data extraction processing to generate second green light channel data; and the second red light channel data and the second green light channel data are respectively subjected to remote photoplethysmography signal bandpass.
  • filtering processing to generate second red light filtering data and second green light filtering data; then performing remote photoplethysmography signal noise reduction processing on the second red light filtering data and the second green light filtering data, respectively, Second red light signal data and second green light signal data are generated.
  • the first main control module is specifically configured to, when the light source information is the red light, intercept the latest data segment whose length is the display time length from the first red light signal data, and generate the first data segment. a red light display data; then perform waveform image data conversion processing on the first red light display data to generate first red light waveform image data; then send the first red light waveform image data to the display screen Perform the first red light waveform display processing;
  • the first main control module is specifically configured to, when the light source information is the green light, intercept the latest data segment whose length is the display time length from the first green light signal data, and generate the first data segment. a green light display data; then perform waveform image data conversion processing on the first green light display data to generate first green light waveform image data; then send the first green light waveform image data to the display screen Perform the first green light waveform display processing;
  • the first main control module is specifically configured to intercept the latest data segment whose length is the display time length from the second red light signal data when the light source information is the red and green light, and generate second red light display data, and then perform waveform image data conversion processing on the second red light display data to generate second red light waveform image data; from the second green light signal data, intercept the latest, length For the data segment of the display time length, generate second green light display data, and then perform waveform image data conversion processing on the second green light display data to generate second green light waveform image data;
  • the second red light waveform image data is sent to the display screen for second red light waveform display processing, and the second green light waveform image data is sent to the display screen for second green light waveform display processing.
  • the first protocol includes Hypertext Transfer Protocol HTTP and Hypertext Transfer Security Protocol HTTPS;
  • the second signal data is equal to the first signal data; the second device token information is equal to the first device token information; the second device type information is equal to the first device type information; the The second age information is equal to the first age information; the second gender information is equal to the first gender information; the second height information is equal to the first height information; the second weight information is equal to the first a weight information;
  • the first communication module is specifically used for accessing the Internet through a mobile communication network, a wireless local area network or a wired local area network;
  • the second communication module is specifically used for accessing the Internet through a mobile communication network, a wireless local area network or a wired local area network.
  • the legality verification module is specifically configured to query the legal token list according to the second device token information, and when the second device token information satisfies the legal token list, the legality The verification process was successful.
  • the parameter verification module is specifically configured to check the second signal data, the second device type information, the second age information, the second gender information, the second height information and the second Whether all the weight information is not empty, when the second signal data, the second device type information, the second age information, the second gender information, the second height information and the second weight When all the information is not empty, the parameter verification process is successful.
  • the data preprocessing module includes a plurality of sub-preprocessing modules;
  • the artificial intelligence blood pressure prediction module includes a plurality of sub-blood pressure prediction models;
  • the data preprocessing module is specifically configured to select the corresponding first sub-preprocessing module according to the prediction model identification information, and perform the analysis on the second signal data, the second age information, the second gender information, the first Second, the height information and the second weight information are used to prepare the input data of the first sub-blood pressure prediction model to generate the first model input data;
  • the first sub-preprocessing module is specifically configured to perform baseline drift elimination processing on the second signal data to generate first process signal data, and then perform noise reduction processing on the first process signal data to generate a second process signal Then, standard sampling and normalization are performed on the second process signal data to generate standard signal data; and then according to the input data format requirements of the first sub-blood pressure prediction model, the second signal data, the first sub-blood pressure prediction model Two age information, the second gender information, the second height information, the second weight information and the standard signal data are packaged into the first model input data;
  • the artificial intelligence blood pressure prediction module is specifically configured to select a corresponding first sub-blood pressure prediction model according to the prediction model identification information, input data to the first model, perform a first blood pressure prediction calculation process, and generate the diastolic blood pressure. data and the systolic blood pressure data.
  • a system for blood pressure prediction based on video data includes a first device and a cloud server.
  • the first device uses the first camera to shoot video of the epidermal area of the test object to obtain first video data, and then performs remote photoplethysmography signal data conversion processing on the first video data to obtain first signal data; Perform heart rate analysis on the first signal data to obtain heart rate data, and use the artificial intelligence blood pressure prediction model to input data into the model that combines the test subject's personality data (age, gender, height, weight, etc.) and the first signal data, and perform blood pressure analysis.
  • the cloud server returns the analysis data (heart rate data, systolic blood pressure data, diastolic blood pressure data) to the first device for display.
  • the test object does not need to wear any specific collection device, and the real-time detection or continuous monitoring of the test object can be started at any time through the first device, which reduces the difficulty of real-time blood pressure detection or continuous monitoring of the test object and enriches the light volume.
  • FIG. 1 is a schematic diagram of a system for blood pressure prediction based on video data provided by an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a photographing method provided by an embodiment of the present invention.
  • a system for blood pressure prediction based on video data includes a first device and a cloud server; wherein the first device includes a first main control module, a first camera, a display screen, and a first communication module; the cloud The server includes a second main control module, a legality verification module, a parameter verification module, a data preprocessing module, an artificial intelligence blood pressure prediction module and a second communication module.
  • the first device here may be a terminal device, a computer, a notebook, a mobile phone, a tablet computer, a mobile phone, a computer, a laptop, a mobile phone, a tablet computer, a mobile communication network, a local area network (wired or wireless access method) or a wide area network (wired or wireless access method) that can access the Internet.
  • terminal or server etc.
  • the cloud server here can be a terminal device, an independent server, a virtual server or a server based on cloud architecture that can access the Internet through a mobile communication network, a local area network (wired or wireless access) or a wide area network (wired or wireless access). .
  • the first device uses a first camera to capture video of the epidermal region of the test object to obtain first video data, and then performs remote optical imaging on the first video data.
  • the plethysmograph signal data is converted and processed to obtain the first signal data;
  • the cloud server performs heart rate analysis on the first signal data to obtain the heart rate data, and uses the artificial intelligence blood pressure prediction model to combine the test subject's personality data (age, gender, height, information such as body weight) and the model input data of the first signal data, perform blood pressure analysis to obtain blood pressure data (systolic blood pressure, diastolic blood pressure); finally, the cloud server returns the analysis data (heart rate data, systolic blood pressure data, diastolic blood pressure data) to the first a device to display.
  • FIG. 1 is a schematic diagram of a system for predicting blood pressure based on video data provided by an embodiment of the present invention.
  • the system for predicting blood pressure based on video data provided by an embodiment of the present invention includes: a first device 1 and a cloud server 2.
  • the first device 1 includes a first main control module 11 , a first camera 12 , a lighting module 13 , a display screen 14 and a first communication module 15 .
  • the first main control module 11 is the control unit of the mobile phone
  • the first camera 12 is the main camera of the mobile phone
  • the light module 13 is the flashlight of the mobile phone
  • the display screen 14 is the screen of the mobile phone
  • the first camera 12 is the main camera of the mobile phone.
  • a communication module 15 is a mobile phone communication unit.
  • the first main control module 11 is used for invoking the first camera 12 and the lighting module 13 to perform photographing processing on the epidermal area of the test object for a first duration, thereby generating first video data.
  • the epidermal area is a preset skin area of the test object
  • the first duration is a preset continuous shooting time
  • the first main control module 11 is specifically configured to call the lighting module 13 to illuminate the epidermis area of the test object, and after the lens of the first camera 12 covers the epidermis area, use the A camera 12 performs photographing processing on the epidermis area for a first duration, thereby generating first video data.
  • the epidermis area is the skin area of the fingertip
  • the first duration is 26 seconds.
  • FIG. 2 which is a schematic diagram of the shooting method provided by the embodiment of the present invention
  • the test object uses a flash to illuminate the skin of the fingertip, and at the same time the main camera It completely covers the skin of the fingertip, and the mobile phone control unit uses the main camera to capture the skin of the fingertip and obtains a piece of first video data with a length of 26 seconds.
  • the display screen 14 is used for receiving the first video data sent by the main control module and performing playback processing.
  • the first video data that is being shot will be played on the display screen in real time.
  • the test subject should stick the main camera close to the skin of the fingertip, so the first video screen being displayed should be full red.
  • the first main control module 11 is further configured to perform light source channel data extraction processing on the first video data according to the light source information to generate first channel data; and then perform remote photoplethysmography signal data conversion processing on the first channel data to generate The first signal data; and then send the first signal data to the display screen 14 according to the display time length for signal waveform display processing.
  • the light source information includes red light, green light, and red-green light.
  • the conventional photoplethysmography signal can only be collected under the condition of red light or infrared light source; the embodiment of the present invention supports shooting in a common light source environment (such as a sunlight environment), not only supports the identification of red light signal conversion into remote photoplethysmography It also supports the simultaneous identification of red and green light signals.
  • a common light source environment such as a sunlight environment
  • the first main control module 11 is specifically configured to perform red light channel data extraction processing on the first video data when the light source information is red light to generate first channel data; When the light source information is green light, the green light channel data extraction process is performed on the first video data to generate the first channel data; when the light source information is red and green light, the red light channel data extraction process is performed on the first video data to generate the first channel data. Red light channel data, performing green light channel data extraction processing on the first video data to generate first green light channel data, and encapsulating the first red light channel data and the first green light channel data into first channel data.
  • the first main control module 11 is specifically configured to perform red light channel data extraction processing on the first video data when the light source information is red light to generate the first channel data, and the specific implementation manner is as follows:
  • the first main control module 11 is specifically configured to perform frame image extraction processing on the first video data when the light source information is red light to obtain a plurality of first frame image data; Count the number of first red pixels that meet the threshold range of red light pixels, generate a first total number, and calculate the sum of pixel values for all first red pixels, generate the sum of the first pixel values, and then sum the first pixel values.
  • the ratio to the first total number is used as the first frame of red light channel data corresponding to each first frame of image data; then all the first frame of red light channel data are arranged in chronological order to generate first channel data.
  • each first frame image is a two-dimensional bitmap, which is composed of multiple pixels, each pixel has a corresponding pixel value, and the first frame image data is all the pixels in the first frame image.
  • the pixel value set of ; the red pixel threshold range is a range of pixel values, and the pixels whose pixel values conform to this range are identified as red pixels.
  • 24 frame images can be extracted from 1 second of video data.
  • the number of a red pixel is the first total number NR
  • the sum of the pixel values of all the first red pixels is the sum of the first pixel values SR
  • the first first frame red light channel data CR SR /N R , here if the frame red light channel data must be an integer, it needs to be rounded; and so on, and finally get 624 first frame red light channel data
  • the first channel data in chronological order is ⁇ C R1 , C R2 ... C R624 ⁇ .
  • the green light channel data extraction processing is performed on the first video data, and the first channel data is generated, and the specific implementation manner is as follows:
  • the first main control module 11 is specifically configured to perform frame image extraction processing on the first video data when the light source information is green light to obtain a plurality of second frame image data; Count the number of first green pixels that meet the green pixel threshold range, generate a second total, and calculate the sum of pixel values for all the first green pixels to generate a second sum of pixel values, and then sum the second pixel values.
  • the ratio to the second total number is used as the first frame of green light channel data corresponding to each second frame of image data; then all the first frame of green light channel data are arranged in chronological order to generate first channel data.
  • each second frame image is a two-dimensional bitmap, which is composed of multiple pixel points, each pixel point has a corresponding pixel value, and the second frame image data is all the pixel points in the second frame image
  • the green pixel threshold range is a range of pixel values, and the pixels whose pixel values conform to this range are identified as green pixels.
  • 24 frame images can be extracted from 1 second of video data.
  • the red light channel data extraction process is performed on the first video data to generate the first red light channel data
  • the green light channel data extraction process is performed on the first video data to generate the first green light channel data
  • the first main control module 11 is specifically configured to perform video frame image extraction processing on the first video data when the light source information is red and green light to obtain a plurality of third frame image data; in each third frame image data, The number of second red pixels whose pixel value meets the threshold range of red light pixels is counted to generate a third total number, and the sum of pixel values for all second red pixels is calculated to generate a third pixel value sum, and then the third pixel value is calculated.
  • the ratio of the sum of the values to the third total is used as the second frame of red light channel data corresponding to each third frame of image data, and then for all the second frames of red light channel data, chronologically arranged to generate the first red light channel data.
  • the frame green light channel data is arranged in chronological order to generate the first green light channel data; the first red light channel data and the first green light channel data are then subjected to multi-channel data encapsulation processing to generate the first channel data.
  • 24 frame images can be extracted from 1 second of video data.
  • the number of the second red pixels is the third total number N' R
  • the number of the second green pixels is the fourth total number N' G
  • the sum of the pixel values of all the second red pixels is the third pixel value sum S' R
  • all The pixel value summation of the second green pixel point is the fourth pixel value summation S'G ;
  • the first main control module 11 is specifically configured to perform remote photoplethysmographic signal bandpass filtering processing on the first channel data when the light source information is red light, and generate a first The red light filtered data is then subjected to remote photoplethysmographic signal noise reduction processing on the first red light filtered data to generate first red light signal data.
  • the control unit of the mobile phone presets a frequency threshold range for the bandpass filtering of the remote photoplethysmography signal, the control unit regards the first channel data as a segment signal data, and calculate the signal frequency of each signal data point in the first channel data, and then based on the band-pass filtering principle, the red light channel corresponding to the signal data in the first channel data whose signal frequency is lower or higher than the frequency threshold range Delete the data to obtain the first red light filter data;
  • the common band-pass filter frequency threshold range is 0.5 Hz to 10 Hz; when band-pass filtering is performed on some mobile phones, considering the limited processing capability of the mobile phone, the frequency threshold is used here.
  • Finite Impulse Response (FIR) filter module can also be used; the process of performing remote photoplethysmography signal noise reduction processing on the first red light filter data is similar to the bandpass filter processing process, which can be regarded as After secondary filtering, the first signal data, which is specifically the first red light signal data, is finally obtained.
  • FIR Finite Impulse Response
  • the first main control module 11 is specifically configured to perform remote photoplethysmography signal bandpass filtering processing on the first channel data when the light source information is green light to generate the first green light filtering data, and then performing remote photoplethysmographic signal noise reduction processing on the first green light filtering data to generate first green light signal data.
  • the light source information is green
  • the first channel data is composed of green light channel data
  • the control unit of the mobile phone regards the first channel data as a piece of signal data, and calculates the signal frequency of each signal data point in the first channel data, and then based on Band-pass filtering principle, delete the green light channel data corresponding to the signal data in the first channel data whose signal frequency is lower than or higher than the remote photoplethysmograph signal band-pass filtering frequency threshold range, and obtain the first green light filtering data; Then, the remote photoplethysmography signal noise reduction process is performed on the first red light filtered data, and finally first signal data, specifically the first green light signal data, is obtained.
  • the first main control module 11 is specifically configured to perform red light channel data extraction processing on the first channel data to generate a second red light channel when the light source information is red and green light perform green channel data extraction processing on the first channel data to generate second green channel data; and then perform remote photoplethysmography signal bandpass on the second red channel data and the second green channel data respectively.
  • the filtering process generates the second red light filtering data and the second green light filtering data; then the second red light filtering data and the second green light filtering data are respectively subjected to remote photoplethysmographic signal noise reduction processing to generate the second red light filtering data and the second green light filtering data. light signal data and second green light signal data.
  • the light source information is green
  • the first channel data is composed of red light channel data and green light channel data
  • the control unit of the mobile phone filters and denoises the number of red light channels respectively to obtain the second red light signal data and the second green light signal data
  • the finally obtained first signal data is specifically composed of the second red light signal data and the second green light signal data.
  • the first main control module 11 is specifically configured to intercept the latest data whose length is the display time length from the first red light signal data when the light source information is red light. segment, generate the first red light display data; then perform waveform image data conversion processing on the first red light display data to generate the first red light waveform image data; then send the first red light waveform image data to the display screen 14 for processing The first red light waveform display processing.
  • the display duration is the duration of the latest waveform for display.
  • the control unit of the mobile phone will intercept the latest signal data with a duration of 1 second from the first red light signal data.
  • the corresponding first red light waveform image data is a display waveform with a length of 1 second, and the waveform will be displayed on the screen of the mobile phone.
  • the waveform color can be set to red when displayed.
  • the first main control module 11 is specifically configured to intercept the latest data whose length is the display time length from the first green light signal data when the light source information is green light segment, generate the first green light display data; then perform waveform image data conversion processing on the first green light display data to generate the first green light waveform image data; then send the first green light waveform image data to the display screen 14 for processing The first green light waveform is displayed for processing.
  • the control unit of the mobile phone will intercept the latest signal data with a duration of 1 second from the first green light signal data
  • the corresponding first green light waveform image data is a display waveform with a length of 1 second, and the waveform will be displayed on the screen of the mobile phone.
  • the waveform color can be set to green when displayed.
  • the first main control module 11 is specifically configured to, when the light source information is red and green light, intercept the latest data whose length is the display time length from the second red light signal data. data segment, generate second red light display data, and then perform waveform image data conversion processing on the second red light display data to generate second red light waveform image data; from the second green light signal data, intercept the latest, length
  • the second green light display data is generated, and the waveform image data conversion processing is performed on the second green light display data to generate the second green light waveform image data; then the second red light waveform image data is It is sent to the display screen 14 for the second red light waveform display processing, and the second green light waveform image data is sent to the display screen 14 for the second green light waveform display processing.
  • the control unit of the mobile phone will read the data from the second red light signal data and the second green light signal data respectively.
  • the latest signal data with a duration of 1 second is intercepted as the second red light display data and the second green light display data.
  • the second red light display data and the second green light display data are respectively It is a display waveform with a length of 1 second. These two waveforms will be displayed on the screen of the mobile phone. For the convenience of distinguishing, here you can set the display waveform of the second red light display data to red when displaying, and set the second red light display data to red.
  • the display waveform of green light display data is set to green.
  • the first main control module 11 is further configured to store the first signal data, the first device token information, the first device type information, the first age information, the first gender information, the first height information and the first weight information of the test subject. , encapsulated into a first data packet according to the first protocol, and send the first data packet to the cloud server 2 by using the first communication module 15 .
  • the first device token information is legal device token information allocated to the first device
  • the first device type information is the specific device type information of the first device
  • the information and the first weight information are the age, gender, height and weight information of the test subject respectively
  • the first protocol is Hyper Text Transfer Protocol (HTTP) or Hyper Text Transfer Protocol (Hyper Text Transfer Protocol over Secure Socket) Layer, HTTPS).
  • the first communication module 15 is specifically used for accessing the Internet through a mobile communication network, a wireless local area network or a wired local area network.
  • the cloud server 2 includes a second main control module 21 , a legality verification module 22 , a parameter verification module 23 , a data preprocessing module 24 , an artificial intelligence blood pressure prediction module 25 and a second communication module 26 .
  • the cloud server 2 is a cloud platform that uses the .NET kernel to superimpose the Tensorflow framework
  • the second main control module 21 is the management unit of the cloud platform
  • the legality verification module 22 is the first business processing unit of the cloud platform.
  • the verification module 23 is the second business processing unit of the cloud platform
  • the data preprocessing module 24 is the data processing unit of the cloud platform
  • the artificial intelligence blood pressure prediction module 25 is the computing unit of the cloud platform
  • the second communication module 26 is specifically the server of the cloud platform Communication processing unit.
  • the .NET kernel is a general open-source development platform, which is developed and provided by Microsoft, and can be adapted to processors of various architectures (such as the x86 and x64 architectures of Intel Corporation of the United States, ARM32 and ARM32 of ARM Corporation of the United Kingdom).
  • ARM64 architecture can adapt to a variety of operating systems (such as Windows operating system Windows, Apple computer operating system macOS and user network operating system Linux);
  • TensorFlow is an open source artificial intelligence model learning framework, which is developed and provided by Google. , which has rich applications in graphics classification, audio processing, recommendation systems, and natural language processing, and is currently the mainstream artificial intelligence model learning framework.
  • the second main control module 21 is configured to perform data analysis processing on the first data packet according to the first protocol to obtain second signal data, second device token information, second device type information, second age information, and second gender information , second height information and second weight information.
  • the second signal data is equal to the first signal data;
  • the second device token information is equal to the first device token information;
  • the second device type information is equal to the first device type information;
  • the second age information is equal to the first age information;
  • the second gender information is equal to the first gender information;
  • the second height information is equal to the first height information;
  • the second weight information is equal to the first weight information.
  • the legality verification module 22 is configured to perform legality verification processing on the token information of the second device according to the legal token list.
  • the valid token list is a vector table that stores all valid device token information.
  • the legality verification module 22 is specifically configured to query the legal token list according to the second device token information, and when the second device token information satisfies the legal token list, The legality verification process is successful.
  • the first service processing unit obtains the legal token list from the database, and then polls all legal device token information in the legal token list, when it finds that the second device token information exists in the legal token list , the legality verification process is successful.
  • the parameter verification module 23 is configured to perform parameters on the second signal data, the second device type information, the second age information, the second gender information, the second height information and the second weight information when the legality verification process is successful. Check processing.
  • the parameter verification module 23 is specifically configured to check the second signal data, the second device type information, the second age information, the second gender information, the second height information, and the second Whether all the weight information is not empty, when the second signal data, the second device type information, the second age information, the second gender information, the second height information and the second weight information are all not empty, the parameter verification process is successful .
  • the second service processing unit checks whether the second signal data, the second device type information, the second age information, the second gender information, the second height information, and the second weight information are all not empty, when the second signal When none of the data, the second device type information, the second age information, the second gender information, the second height information, and the second weight information are empty, the parameter integrity verification process is successful.
  • the second main control module 21 is further configured to perform heart rate calculation processing according to the second signal data to generate heart rate data when the parameter integrity verification processing is successful.
  • the second signal data is signal data with a length of 26 seconds.
  • the management unit regards the second signal data as continuous waveform data, and sequentially extracts the time points of the wave peaks of the continuous waveform data as the signal time points, and uses the adjacent signal time points as the signal time points.
  • the data preprocessing module 24 is used to prepare the input data of the blood pressure prediction model for the second signal data, the second age information, the second gender information, the second height information and the second weight information according to the prediction model identification information, and generate a model Input data.
  • the prediction model identification information includes at least a first convolutional neural network model identification and a second convolutional neural network model identification;
  • the data preprocessing module 24 includes a plurality of sub-preprocessing modules.
  • the data preprocessing module 24 is specifically configured to select the corresponding first sub-preprocessing module according to the prediction model identification information, and perform the analysis on the second signal data, the second age information, the second The gender information, the second height information and the second weight information are processed to prepare the input data of the first sub-blood pressure prediction model to generate the first model input data.
  • the data processing unit selects the first sub-preprocessing module, which is specifically the preprocessing module of the first convolutional neural network model.
  • the first sub-preprocessing module which is specifically the preprocessing module of the first convolutional neural network model.
  • information, the second gender information, the second height information, and the second weight information prepare the input data of the first convolutional neural network model, and generate the first model input data, which is specifically the input data of the first convolutional neural network model.
  • the data processing unit selects the first sub-preprocessing module that is specifically the second convolutional neural network model preprocessing module, and the second signal data, the second The age information, the second gender information, the second height information, and the second weight information are processed to prepare the input data of the second convolutional neural network model, and the first model input data, which is specifically the input data of the second convolutional neural network model, is generated.
  • the first sub-preprocessing module is specifically configured to perform baseline drift elimination processing on the second signal data, generate first process signal data, and then perform noise reduction on the first process signal data processing to generate second process signal data, then standard sampling and normalization processing are performed on the second process signal data to generate standard signal data; and then according to the input data format requirements of the first sub-blood pressure prediction model, the second signal data, The second age information, the second gender information, the second height information, the second weight information and the standard signal data are packaged into the first model input data.
  • the process of eliminating baseline drift can use multiple fitting filtering, median filtering, infinite impulse response (Infinite Impulse Response, IIR) filtering, fast Fourier transform (fast Fourier transform, FFT) ) filtering or wavelet transform filtering and other processing methods; noise reduction processing can use FFT filtering, band-pass filtering or band-stop filtering and other processing methods; normalization processing commonly uses linear normalization; A standard sampling of the signal data is required, and the sampling frequency of the standard sampling is related to the input data format requirements of the blood pressure prediction model corresponding to the preprocessing sub-module.
  • IIR Infinite Impulse Response
  • FFT fast Fourier transform
  • the prediction model identification information is the first convolutional neural network model identification
  • the first sub-preprocessing module is the first convolutional neural network model preprocessing module
  • the first sub-preprocessing module uses multinomial fitting filtering, median Filter, infinite impulse response filtering, fast Fourier transform filtering or wavelet transform filtering and other processing means to eliminate the baseline drift of the second signal data to generate the first process signal data
  • fast Fourier transform filtering, bandpass filtering or Band-stop filtering and other processing means perform noise reduction processing on the first process signal data to generate second process signal data
  • use the sampling frequency corresponding to the first convolutional neural network model to perform standard sampling on the second process signal data
  • linear The normalization method normalizes the sampled data to generate standard signal data
  • the second height information, the second weight information and the standard signal data are packaged into first model input data, which is specifically the first
  • the prediction model identification information is the second convolutional neural network model identification
  • the first sub-preprocessing module is the second convolutional neural network model preprocessing module
  • the first sub-preprocessing module uses multinomial fitting filtering, middle Value filtering, infinite impulse response filtering, fast Fourier transform filtering or wavelet transform filtering and other processing means to eliminate baseline drift processing on the second signal data to generate the first process signal data; then use fast Fourier transform filtering, bandpass filtering or band-stop filtering and other processing means to perform noise reduction processing on the first process signal data to generate second process signal data; then use the sampling frequency corresponding to the second convolutional neural network model to perform standard sampling on the second process signal data, and then pass
  • the linear normalization method normalizes the sampled data to generate standard signal data; finally, according to the input data format requirements of the second convolutional neural network model, the second signal data, second age information, second gender information, The second height information, the second weight information, and the standard signal data are packaged into first model input data that is specifically the second
  • the artificial intelligence blood pressure prediction module 25 is used for inputting data to the model according to the identification information of the prediction model, performing blood pressure prediction calculation processing, and generating diastolic blood pressure data and systolic blood pressure data.
  • the artificial intelligence blood pressure prediction module 25 includes a plurality of sub blood pressure prediction models.
  • the artificial intelligence blood pressure prediction module 25 provides at least two blood pressure prediction models: a first convolutional neural network model and a second convolutional neural network model:
  • the first convolutional neural network model including multi-layer convolutional neural network layers and fully connected layers, each convolutional neural network layer includes 1 convolutional layer and 1 pooling layer; the convolutional layer is responsible for the model
  • the input data is used for blood pressure feature extraction calculation, and the pooling layer is to downsample the extraction results of the convolutional layer.
  • the output of each convolutional neural network layer is used as the input of the next convolutional neural network layer.
  • the calculation results of the integrated neural network layer are input to the fully connected layer for regression calculation, and the systolic blood pressure data and the diastolic blood pressure data are obtained;
  • the second convolutional neural network model including a two-dimensional convolutional layer, a maximum pooling layer, a batch normalization layer, an activation layer, an addition layer, a global average pooling layer, a random drop layer and a fully connected layer.
  • the 2D convolution layer can contain multiple sub-convolution layers, which are responsible for performing multiple convolution calculations on the input data of the model.
  • the convolution result output by the 2D convolution layer contains multiple 1D vectors, and the maximum pooling layer is in each 1D vector.
  • the convolution result is sampled by taking the maximum value in the vector to reduce the amount of data.
  • the batch normalization layer performs data uniformity processing on the output results of the maximum pooling layer, and the activation layer uses a nonlinear activation function.
  • the output results of the batch normalization layer are connected by a neural network, the addition layer performs weighted addition calculation on the output results of the activation layer, the global average pooling layer performs a weighted average calculation of the whole data on the output results of the addition layer, and the random discard layer is calculated according to randomness.
  • the output results of the global average pooling layer are clipped, and finally the fully connected layer is used to perform binary regression on the clipped random drop layer output results to calculate the output diastolic blood pressure data and systolic blood pressure data.
  • the artificial intelligence blood pressure prediction module 25 is specifically configured to select a corresponding first sub-blood pressure prediction model according to the prediction model identification information, input data to the first model, and perform a first blood pressure prediction The arithmetic processing generates diastolic blood pressure data and systolic blood pressure data.
  • the prediction model identification information is the identification of the first convolutional neural network model
  • the input data of the first model is the input data of the first convolutional neural network model
  • the calculation unit selects the first sub-blood pressure prediction of the first convolutional neural network model.
  • the model is to perform the first convolutional neural network blood pressure prediction operation processing on the input data of the first convolutional neural network model to generate diastolic blood pressure data and systolic blood pressure data.
  • the prediction model identification information is the second convolutional neural network model identification
  • the first model input data is the second convolutional neural network model input data
  • the calculation unit selects the first sub-blood pressure specifically of the second convolutional neural network model.
  • input data to the second convolutional neural network model perform the second convolutional neural network blood pressure prediction operation processing, and generate diastolic blood pressure data and systolic blood pressure data.
  • the second main control module 21 is also used to set the status code data as normal status code information, and then form return data according to the heart rate data, diastolic blood pressure data and systolic blood pressure data, and then encapsulate the return data and the status code data into a form according to the first protocol.
  • the second data packet is sent to the first device by using the second communication module 26 .
  • the second communication module 26 is specifically configured to access the Internet through a mobile communication network, a wireless local area network, or a wired local area network.
  • the management unit of the cloud platform will set the status code data as normal status code information, and set the returned data by the heart rate data + diastolic blood pressure data + systolic blood pressure data, and package the returned data and status code data according to the HTTP or HTTPS data packaging format to generate a second data package; and use the communication processing unit of the cloud platform to send the second data package to the mobile phone .
  • the first main control module 11 is further configured to perform data analysis processing on the second data packet according to the first protocol to obtain return data and status code data; when the status code data is normal status code information, obtain heart rate data, Diastolic blood pressure data and systolic blood pressure data; and then send the heart rate data, diastolic blood pressure data and systolic blood pressure data to the display screen 14 for heart rate and blood pressure data display processing.
  • the control unit of the mobile phone obtains that the heart rate data is 76 beats/min, the diastolic blood pressure data is 85 mmHg, and the systolic blood pressure data is 112 mmHg; After the data is sent, the following information is displayed: "Heart rate: 76 bpm”, "Diastolic blood pressure: 85 mmHg” and "Systolic blood pressure: 112 mmHg”.
  • a system for blood pressure prediction based on video data includes a first device and a cloud server.
  • the first device uses the first camera to shoot video of the epidermal area of the test object to obtain first video data, and then performs remote photoplethysmography signal data conversion processing on the first video data to obtain the first signal data;
  • Heart rate analysis is performed on the first signal data to obtain heart rate data, and an artificial intelligence blood pressure prediction model is used to input data into the model that combines the test subject's personality data (age, gender, height, weight, etc.) and the first signal data to perform blood pressure analysis.
  • the cloud server returns the analysis data (heart rate data, systolic blood pressure data, diastolic blood pressure data) to the first device for display.
  • the test object does not need to wear any specific collection device, and the real-time detection or continuous monitoring of the test object can be started at any time through the first device, which reduces the difficulty of real-time blood pressure detection or continuous monitoring of the test object and enriches the light volume.
  • a software module can be placed in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.
  • RAM random access memory
  • ROM read only memory
  • electrically programmable ROM electrically erasable programmable ROM
  • registers hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Signal Processing (AREA)
  • Animal Behavior & Ethology (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Public Health (AREA)
  • Cardiology (AREA)
  • Computer Security & Cryptography (AREA)
  • Physiology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychiatry (AREA)
  • Vascular Medicine (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)

Abstract

本发明实施例涉及一种基于视频数据进行血压预测的系统，所述基于视频数据进行血压预测的系统包括：第一设备和云服务器；其中，第一设备包括第一主控模块、第一摄像头、灯光模块、显示屏和第一通讯模块；云服务器包括第二主控模块、合法性校验模块、参数校验模块、数据预处理模块、人工智能血压预测模块和第二通讯模块。使用本发明实施例提供的基于视频数据进行血压预测的系统，无需佩戴任何特定的采集装置即可对血压进行实时检测和持续监测，降低了对测试对象进行实时检测、监测的实现难度，丰富了光体积描计法在监护领域的应用场景。

Description

一种基于视频数据进行血压预测的系统
本申请要求于2020年10月12日提交中国专利局、申请号为202011086786.5、发明名称为“一种基于视频数据进行血压预测的系统”的中国专利申请的优先权。
技术领域
本发明涉及信号处理技术领域，特别涉及一种基于视频数据进行血压预测的系统。
背景技术
光体积描计法，是借助光电手段在活体组织中检测血液容积变化的一种无创检测方法。心脏搏动会使得血管内单位面积的血流量形成周期性变化，与之对应的血液体积也相应发生变化，从而使得反映血液吸收光量的光体积描计信号也会发生周期性变化，常规光体积描计信号的周期性变化与心脏搏动、血压变化是密切相关的。对常规光体积描计信号进行心搏间期分析，可以获得心脏的心率数据；对常规光体积描计信号使用训练成熟的人工智能血压预测模型进行血压相关分析，可以获得血压的舒张压和收缩压数据。远程光体积描计信号，是皮肤对光吸收或反射产生的周期性信号。基于远程光体积描计信号进行心率计算和血压分析的方法与基于光体积描计信号的方法是一样的。但对常规光体积描计信号的采集，需要测试对象佩戴定制的采集装置(例如，指夹、耳夹等采集设备)，这既不便于测试对象对自身的即时检测，也不便于对测试对象进行持续监测。而要获得远程光体积描计信号，只需通过高清摄像头对皮肤表面进行拍摄，并对视频数据做进一步光通道数据转换即可。
发明内容
本发明的目的，就是针对现有技术的缺陷，本发明提供的一种基于视频数据进行血压预测的系统，包括第一设备与云服务器；使用该系统，无需佩戴任何特定的采集装置即可通过第一设备启动实时检测和持续监测，既能降低对测试对象进行血压实时检测和持续监测的实现难度，又可以丰富光体积描计法在监护领域的应用场景。
为实现上述目的，本发明提供了一种基于视频数据进行血压预测的系统，包括：第一设备和云服务器；
所述第一设备包括第一主控模块、第一摄像头、灯光模块、显示屏和第一通讯模块；
所述第一主控模块用于调用所述第一摄像头和所述灯光模块，对测试对象的表皮区域进行第一持续时间的拍摄处理，从而生成第一视频数据；
所述显示屏用于接收所述主控模块发送的所述第一视频数据，进行播放处理；
所述第一主控模块还用于根据光源信息，对所述第一视频数据进行光源通道数据提取处理，生成第一通道数据；然后对所述第一通道数据，进行远程光体积描计信号数据转换处理，生成第一信号数据；再根据显示时间长度将所述第一信号数据发送给所述显示屏进行信号波形显示处理；
所述第一主控模块还用于将所述第一信号数据、第一设备令牌信息、第一设备类型信息、所述测试对象的第一年龄信息、第一性别信息、第一身高信息和第一体重信息，根据第一协议封装成第一数据包，并利用所述第一通讯模块将所述第一数据包发送给所述云服务器；
所述云服务器包括第二主控模块、合法性校验模块、参数校验模块、数据预处理模块、人工智能血压预测模块和第二通讯模块；
所述第二主控模块用于将所述第一数据包按所述第一协议进行数据解析 处理，得到第二信号数据、第二设备令牌信息、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息；
所述合法性校验模块用于根据合法令牌列表，对所述第二设备令牌信息，进行合法性校验处理；
所述参数校验模块用于当所述合法性校验处理成功时，对所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息，进行参数校验处理；
所述第二主控模块还用于当所述参数完整性校验处理成功时，根据所述第二信号数据，进行心率计算处理，生成心率数据；
所述数据预处理模块用于根据预测模型标识信息，对所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息，进行血压预测模型输入数据准备处理，生成模型输入数据；
所述人工智能血压预测模块用于根据所述预测模型标识信息，对所述模型输入数据，进行血压预测运算处理，生成舒张压数据和收缩压数据；
所述第二主控模块还用于设置状态码数据为正常状态码信息，然后根据所述心率数据、所述舒张压数据和所述收缩压数据组成返回数据，再将所述返回数据和所述状态码数据，根据所述第一协议封装成第二数据包，并利用所述第二通讯模块将所述第二数据包发送给所述第一设备；
所述第一主控模块还用于将所述第二数据包按所述第一协议进行数据解析处理，得到所述返回数据和所述状态码数据；当所述状态码数据为所述正常状态码信息时、从所述返回数据中获取所述心率数据、所述舒张压数据和所述收缩压数据；再将所述心率数据、所述舒张压数据和所述收缩压数据发送给所述显示屏进行心率及血压数据显示处理。
优选的，
所述第一主控模块具体用于调用所述灯光模块对所述测试对象的表皮区域进行照射，并在所述第一摄像头的镜头覆盖了所述表皮区域之后，利用 所述第一摄像头对所述表皮区域进行所述第一持续时间的拍摄处理，从而生成所述第一视频数据。
优选的，
所述第一主控模块具体用于当所述光源信息为红光时，对所述第一视频数据进行红光通道数据提取处理，生成所述第一通道数据；当所述光源信息为绿光时，对所述第一视频数据进行绿光通道数据提取处理，生成所述第一通道数据；当所述光源信息为红绿光时，对所述第一视频数据进行红光通道数据提取处理生成第一红光通道数据，对所述第一视频数据进行绿光通道数据提取处理生成第一绿光通道数据，并将所述第一红光通道数据和所述第一绿光通道数据封装成所述第一通道数据。
优选的，
所述第一主控模块具体用于当所述光源信息为所述红光时，对所述第一视频数据进行帧图像提取处理，得到多个第一帧图像数据；在每个第一帧图像数据中，对像素值满足红光像素阈值范围的第一红色像素点的数量进行统计，生成第一总数，并对所有所述第一红色像素点进行像素值总和计算，生成第一像素值总和，再将所述第一像素值总和与所述第一总数的比值，做为与所述每个第一帧图像数据对应的第一帧红光通道数据；然后对所有所述第一帧红光通道数据，按时间先后顺序排列，生成所述第一通道数据；
所述第一主控模块具体用于当所述光源信息为所述绿光时，对所述第一视频数据进行帧图像提取处理，得到多个第二帧图像数据；在每个第二帧图像数据中，对像素值满足绿光像素阈值范围的第一绿色像素点的数量进行统计，生成第二总数，并对所有所述第一绿色像素点进行像素值总和计算，生成第二像素值总和，再将所述第二像素值总和与所述第二总数的比值，做为与所述每个第二帧图像数据对应的第一帧绿光通道数据；然后对所有所述第一帧绿光通道数据，按时间先后顺序排列，生成所述第一通道数据；
所述第一主控模块具体用于当所述光源信息为所述红绿光时，对所述第 一视频数据进行视频帧图像提取处理，得到多个第三帧图像数据；在每个第三帧图像数据中，对像素值满足所述红光像素阈值范围的第二红色像素点的数量进行统计，生成第三总数，并对所有所述第二红色像素点进行像素值总和计算，生成第三像素值总和，再将所述第三像素值总和与所述第三总数的比值，做为与所述每个第三帧图像数据对应的第二帧红光通道数据，然后对所有所述第二帧红光通道数据，按时间先后顺序排列，生成所述第一红光通道数据；并在所述每个第三帧图像数据中，对像素值满足所述绿光像素阈值范围的第二绿色像素点的数量进行统计，生成第四总数，并对所有所述第二绿色像素点进行像素值总和计算，生成第四像素值总和，再将所述第四像素值总和与所述第四总数的比值，做为与所述每个第三帧图像数据对应的第二帧绿光通道数据，然后对所有所述第二帧绿光通道数据，按时间先后顺序排列，生成所述第一绿光通道数据；再对所述第一红光通道数据和所述第一绿光通道数据，进行多通道数据封装处理，生成所述第一通道数据。
优选的，
所述第一主控模块具体用于当所述光源信息为所述红光时，对所述第一通道数据进行远程光体积描计信号带通滤波处理，生成第一红光滤波数据，再对所述第一红光滤波数据进行远程光体积描计信号降噪处理，生成第一红光信号数据；
所述第一主控模块具体用于当所述光源信息为所述绿光时，对所述第一通道数据进行远程光体积描计信号带通滤波处理，生成第一绿光滤波数据，再对所述第一绿光滤波数据进行远程光体积描计信号降噪处理，生成第一绿光信号数据；
所述第一主控模块具体用于当所述光源信息为所述红绿光时，对所述第一通道数据进行红光通道数据提取处理，生成第二红光通道数据；并对所述第一通道数据进行绿光通道数据提取处理，生成第二绿光通道数据；再对所述第二红光通道数据和所述第二绿光通道数据，分别进行远程光体积描计信 号带通滤波处理，生成第二红光滤波数据和第二绿光滤波数据；然后对所述第二红光滤波数据和所述第二绿光滤波数据，分别进行远程光体积描计信号降噪处理，生成第二红光信号数据和第二绿光信号数据。
优选的，
所述第一主控模块具体用于当所述光源信息为所述红光时，从所述第一红光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第一红光显示数据；再对所述第一红光显示数据，进行波形图像数据转换处理，生成第一红光波形图像数据；然后将所述第一红光波形图像数据发送给所述显示屏进行第一红光波形显示处理；
所述第一主控模块具体用于当所述光源信息为所述绿光时，从所述第一绿光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第一绿光显示数据；再对所述第一绿光显示数据，进行波形图像数据转换处理，生成第一绿光波形图像数据；然后将所述第一绿光波形图像数据发送给所述显示屏进行第一绿光波形显示处理；
所述第一主控模块具体用于当所述光源信息为所述红绿光时，从所述第二红光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第二红光显示数据，再对所述第二红光显示数据，进行波形图像数据转换处理，生成第二红光波形图像数据；从所述第二绿光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第二绿光显示数据，再对所述第二绿光显示数据，进行波形图像数据转换处理，生成第二绿光波形图像数据；然后将所述第二红光波形图像数据发送给所述显示屏进行第二红光波形显示处理，将所述第二绿光波形图像数据发送给所述显示屏进行第二绿光波形显示处理。
优选的，
所述第一协议包括超文本传输协议HTTP和超文本传输安全协议HTTPS；
所述第二信号数据等于所述第一信号数据；所述第二设备令牌信息等于所述第一设备令牌信息；所述第二设备类型信息等于所述第一设备类型信息；所述第二年龄信息等于所述第一年龄信息；所述第二性别信息等于所述第一性别信息；所述第二身高信息等于所述第一身高信息；所述第二体重信息等于所述第一体重信息；
所述第一通讯模块具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网；
所述第二通讯模块具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网。
优选的，
所述合法性校验模块具体用于根据所述第二设备令牌信息，查询所述合法令牌列表，当所述第二设备令牌信息满足所述合法令牌列表时，所述合法性校验处理成功。
优选的，
所述参数校验模块具体用于检查所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息是否全部不为空，当所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息全部不为空时，所述参数校验处理成功。
优选的，
所述数据预处理模块包括多个子预处理模块；所述人工智能血压预测模块包括多个子血压预测模型；
所述数据预处理模块具体用于根据预测模型标识信息，选择对应的第一子预处理模块，对所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息，进行第一子血压预测模型输入数据准备处理，生成第一模型输入数据；
所述第一子预处理模块具体用于对所述第二信号数据进行消除基线漂移处理，生成第一过程信号数据，再对所述第一过程信号数据进行降噪处理，生成第二过程信号数据，然后对所述第二过程信号数据进行标准采样和归一化处理，生成标准信号数据；再按第一子血压预测模型的输入数据格式要求，将所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息、所述第二体重信息和所述标准信号数据封装成所述第一模型输入数据；
所述人工智能血压预测模块具体用于根据所述预测模型标识信息，选择对应的第一子血压预测模型，对所述第一模型输入数据，进行第一血压预测运算处理，生成所述舒张压数据和所述收缩压数据。
本发明实施例提供的一种基于视频数据进行血压预测的系统，包括第一设备和云服务器。其中，第一设备使用第一摄像头对测试对象的表皮区域进行视频拍摄得到第一视频数据，再对第一视频数据进行远程光体积描计信号数据转换处理得到第一信号数据；云服务器则对第一信号数据进行心率分析得到心率数据，并使用人工智能血压预测模型，对结合了测试对象个性数据(年龄、性别、身高、体重等信息)和第一信号数据的模型输入数据，进行血压分析得到血压数据(收缩压、舒张压)；最后，云服务器将分析数据(心率数据、收缩压数据、舒张压数据)返回到第一设备进行显示。使用该系统，测试对象无需佩戴任何特定的采集装置，通过第一设备可随时启动对测试对象的实时检测或持续监测，降低了对测试对象进行血压实时检测或持续监测的难度，丰富了光体积描计法在监护领域的应用场景。
附图说明
图1为本发明实施例提供的一种基于视频数据进行血压预测的系统示意图；
图2为本发明实施例提供的拍摄方法示意图。
具体实施方式
为了使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明作进一步地详细描述，显然，所描述的实施例仅仅是本发明一部份实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
本发明实施例提供的一种基于视频数据进行血压预测的系统，包括第一设备和云服务器；其中，第一设备包括第一主控模块、第一摄像头、显示屏和第一通讯模块；云服务器包括第二主控模块、合法性校验模块、参数校验模块、数据预处理模块、人工智能血压预测模块和第二通讯模块。
这里的第一设备，可以是能通过移动通信网络、局域网(有线或者无线接入方式)或广域网(有线或者无线接入方式)接入互联网的终端设备、电脑、笔记本、手机、平板电脑、移动终端或服务器等。这里的云服务器，可以是能通过移动通信网络、局域网(有线或者无线接入方式)或广域网(有线或者无线接入方式)接入互联网的终端设备、独立服务器、虚拟服务器或基于云架构的服务器。
在本发明实施例提供的一种基于视频数据进行血压预测的系统中，第一设备使用第一摄像头对测试对象的表皮区域进行视频拍摄得到第一视频数据，再对第一视频数据进行远程光体积描计信号数据转换处理得到第一信号数据；云服务器则对第一信号数据进行心率分析得到心率数据，并使用人工智能血压预测模型，对结合了测试对象个性数据(年龄、性别、身高、体重等信息)和第一信号数据的模型输入数据，进行血压分析得到血压数据(收缩压、舒张压)；最后，云服务器将分析数据(心率数据、收缩压数据、舒张压数据)返回到第一设备进行显示。
图1为本发明实施例提供的一种基于视频数据进行血压预测的系统示 意图，如图1所示，本发明实施例提供的基于视频数据进行血压预测的系统包括：第一设备1和云服务器2。
第一设备1包括第一主控模块11、第一摄像头12、灯光模块13、显示屏14和第一通讯模块15。
例如，第一设备1具体为手机，则第一主控模块11为手机的控制单元，第一摄像头12为手机的主摄像头，灯光模块13为手机的闪光灯，显示屏14为手机的屏幕，第一通讯模块15为手机通讯单元。
第一主控模块11用于调用第一摄像头12和灯光模块13，对测试对象的表皮区域进行第一持续时间的拍摄处理，从而生成第一视频数据。
此处，表皮区域为预设的测试对象的皮肤区域，第一持续时间为预设的持续拍摄时间。
在本实施例提供的一个具体实现方式中，第一主控模块11具体用于调用灯光模块13对测试对象的表皮区域进行照射，并在第一摄像头12的镜头覆盖了表皮区域之后，利用第一摄像头12对表皮区域进行第一持续时间的拍摄处理，从而生成第一视频数据。
例如，表皮区域为指尖皮肤区域，第一持续时间为26秒，则如图2为本发明实施例提供的拍摄方法示意图所示，测试对象使用闪光灯对指尖皮肤进行照射，同时将主摄像头完全覆盖在指尖皮肤上，手机控制单元使用主摄像头对指尖皮肤进行拍摄得到的一条长度为26秒的第一视频数据。
显示屏14用于接收主控模块发送的第一视频数据，进行播放处理。
例如，在26秒的拍摄过程中，显示屏上会实时播放正在拍摄的第一视频数据。常规情况下，测试对象应将主摄像头近距离贴在指尖皮肤上，所以正在显示的第一视频画面应为全红。
第一主控模块11还用于根据光源信息，对第一视频数据进行光源通道数据提取处理，生成第一通道数据；然后对第一通道数据，进行远程光体积描计信号数据转换处理，生成第一信号数据；再根据显示时间长度将第一信 号数据发送给显示屏14进行信号波形显示处理。
此处，光源信息包括红光、绿光和红绿光。
这里，采集常规光体积描计信号只能在红光或红外光源条件下采集；本发明实施例支持在普通光源环境(例如日光环境)下拍摄，不仅支持识别红光信号转换为远程光体积描计信号、识别绿光信号转换为远程光体积描计信号，还支持同时识别红绿光信号。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为红光时，对第一视频数据进行红光通道数据提取处理，生成第一通道数据；当光源信息为绿光时，对第一视频数据进行绿光通道数据提取处理，生成第一通道数据；当光源信息为红绿光时，对第一视频数据进行红光通道数据提取处理生成第一红光通道数据，对第一视频数据进行绿光通道数据提取处理生成第一绿光通道数据，并将第一红光通道数据和第一绿光通道数据封装成第一通道数据。
此处，第一主控模块11具体用于当光源信息为红光时，对第一视频数据进行红光通道数据提取处理，生成第一通道数据，的具体实现方式为：
第一主控模块11具体用于当光源信息为红光时，对第一视频数据进行帧图像提取处理，得到多个第一帧图像数据；在每个第一帧图像数据中，对像素值满足红光像素阈值范围的第一红色像素点的数量进行统计，生成第一总数，并对所有第一红色像素点进行像素值总和计算，生成第一像素值总和，再将第一像素值总和与第一总数的比值，做为与每个第一帧图像数据对应的第一帧红光通道数据；然后对所有第一帧红光通道数据，按时间先后顺序排列，生成第一通道数据。
这里，每个第一帧图像都是一个二维位图，其内部由多个像素点组成，每个像素点有一个对应的像素值，第一帧图像数据就是第一帧图像内所有像素点的像素值集合；红光像素阈值范围是一个像素值范围，像素值符合该范围的像素点，被识别为红色像素点。
例如，1秒视频数据可以提取24张帧图像，从长度为26秒的第一视频数据中，可以提取26*24＝624个第一帧图像数据；第1个第一帧图像数据中，第一红色像素点的数量为第一总数N R，所有第一红色像素点的像素值总和为第一像素值总和S R，则第1个第一帧红光通道数据C R＝S R/N R,这里如果帧红光通道数据必须为整数，则需要对其进行取整计算；依次类推，最后得到624个第一帧红光通道数据，按时间先后顺序排列得到的第一通道数据为{C R1,C R2……C R624}。
此处，当光源信息为绿光时，对第一视频数据进行绿光通道数据提取处理，生成第一通道数据，的具体实现方式为：
第一主控模块11具体用于当光源信息为绿光时，对第一视频数据进行帧图像提取处理，得到多个第二帧图像数据；在每个第二帧图像数据中，对像素值满足绿光像素阈值范围的第一绿色像素点的数量进行统计，生成第二总数，并对所有第一绿色像素点进行像素值总和计算，生成第二像素值总和，再将第二像素值总和与第二总数的比值，做为与每个第二帧图像数据对应的第一帧绿光通道数据；然后对所有第一帧绿光通道数据，按时间先后顺序排列，生成第一通道数据。
这里，每个第二帧图像都是一个二维位图，其内部由多个像素点组成，每个像素点有一个对应的像素值，第二帧图像数据就是第二帧图像内所有像素点的像素值集合；绿光像素阈值范围是一个像素值范围，像素值符合该范围的像素点，被识别为绿色像素点。
例如，1秒视频数据可以提取24张帧图像，从长度为26秒的第一视频数据中，可以提取26*24＝624个第二帧图像数据；第1个第二帧图像数据中，第一绿色像素点的数量为第二总数N G，所有第一绿色像素点的像素值总和为第二像素值总和S G，则第1个第一帧绿光通道数据C G＝S G/N G,这里如果帧绿光通道数据必须为整数，则需要对其进行取整计算；依次类推，最后得到624个第一帧绿光通道数据，按时间先后顺序排列得到的第一通道数据为 {C G1,C G2……C G624}。
此处，当光源信息为红绿光时，对第一视频数据进行红光通道数据提取处理生成第一红光通道数据，对第一视频数据进行绿光通道数据提取处理生成第一绿光通道数据，并将第一红光通道数据和第一绿光通道数据封装成第一通道数据，的具体实现方式为：
第一主控模块11具体用于当光源信息为红绿光时，对第一视频数据进行视频帧图像提取处理，得到多个第三帧图像数据；在每个第三帧图像数据中，对像素值满足红光像素阈值范围的第二红色像素点的数量进行统计，生成第三总数，并对所有第二红色像素点进行像素值总和计算，生成第三像素值总和，再将第三像素值总和与第三总数的比值，做为与每个第三帧图像数据对应的第二帧红光通道数据，然后对所有第二帧红光通道数据，按时间先后顺序排列，生成第一红光通道数据；并在每个第三帧图像数据中，对像素值满足绿光像素阈值范围的第二绿色像素点的数量进行统计，生成第四总数，并对所有第二绿色像素点进行像素值总和计算，生成第四像素值总和，再将第四像素值总和与第四总数的比值，做为与每个第三帧图像数据对应的第二帧绿光通道数据，然后对所有第二帧绿光通道数据，按时间先后顺序排列，生成第一绿光通道数据；再对第一红光通道数据和第一绿光通道数据，进行多通道数据封装处理，生成第一通道数据。
例如，1秒视频数据可以提取24张帧图像，从长度为26秒的第一视频数据中，可以提取26*24＝624个第三帧图像数据；第1个第三帧图像数据中，第二红色像素点的数量为第三总数N' R,第二绿色像素点的数量为第四总数N' G，所有第二红色像素点的像素值总和为第三像素值总和S' R，所有第二绿色像素点的像素值总和为第四像素值总和S' G；则第1个第二帧红光通道数据为C' R＝S' R/N' R,依次类推，第一红光通道数据{C' R1,C' R2……C' R624}；第1个第二帧绿光通道数据为C' G＝S' G/N' G,依次类推，第一绿光通道数据{C' G1,C' G2……C' G624}；最后，第一通道数据中包括了第一红光通道数据与第一 绿光通道数据两组一维数据。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为红光时，对第一通道数据进行远程光体积描计信号带通滤波处理，生成第一红光滤波数据，再对第一红光滤波数据进行远程光体积描计信号降噪处理，生成第一红光信号数据。
例如，光源信息为红色，第一通道数据由红光通道数据组成，手机的控制单元预置了一个远程光体积描计信号带通滤波频率阈值范围，则控制单元将第一通道数据视作一段信号数据，并计算第一通道数据中各信号数据点的信号频率，再基于带通滤波原理，将第一通道数据中信号频率低于或高于该频率阈值范围的信号数据对应的红光通道数据删除，得到第一红光滤波数据；这里，带通滤波频率阈值范围常见的为0.5赫兹到10赫兹；在某些手机上进行带通滤波处理时，考虑到手机的处理能力有限，这里使用有限长单位冲激响应((Finite Impulse Response，FIR)滤波模块也可以；再对第一红光滤波数据进行远程光体积描计信号降噪处理的过程与带通滤波处理过程近似，可视为二次滤波。最终得到具体为第一红光信号数据的第一信号数据。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为绿光时，对第一通道数据进行远程光体积描计信号带通滤波处理，生成第一绿光滤波数据，再对第一绿光滤波数据进行远程光体积描计信号降噪处理，生成第一绿光信号数据。
例如，光源信息为绿色，第一通道数据由绿光通道数据组成，手机的控制单元将第一通道数据视作一段信号数据，并计算第一通道数据中各信号数据点的信号频率，再基于带通滤波原理，将第一通道数据中信号频率低于或高于远程光体积描计信号带通滤波频率阈值范围的信号数据对应的绿光通道数据删除，得到第一绿光滤波数据的；再对第一红光滤波数据进行远程光体积描计信号降噪处理，最终得到具体为第一绿光信号数据的第一信号数据。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为红绿光时，对第一通道数据进行红光通道数据提取处理，生成第二红光通道数据；并对第一通道数据进行绿光通道数据提取处理，生成第二绿光通道数据；再对第二红光通道数据和第二绿光通道数据，分别进行远程光体积描计信号带通滤波处理，生成第二红光滤波数据和第二绿光滤波数据；然后对第二红光滤波数据和第二绿光滤波数据，分别进行远程光体积描计信号降噪处理，生成第二红光信号数据和第二绿光信号数据。
例如，光源信息为绿色，第一通道数据由红光通道数据与绿光通道数据组成，手机的控制单元如前文所述，分别对红光通道数进行滤波、降噪处理得到第二红光信号数据和第二绿光信号数据，最终得到的第一信号数据具体由第二红光信号数据和第二绿光信号数据组成。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为红光时，从第一红光信号数据中，截取最新的、长度为显示时间长度的数据段，生成第一红光显示数据；再对第一红光显示数据，进行波形图像数据转换处理，生成第一红光波形图像数据；然后将第一红光波形图像数据发送给显示屏14进行第一红光波形显示处理。
此处，显示时间长度为用于显示的最新波形的时长。
例如，显示时间长度为1秒，光源信息为红色，第一信号数据为第一红光信号数据，则手机的控制单元会从第一红光信号数据中截取最近一段时长为1秒的信号数据作为第一红光显示数据，对应的第一红光波形图像数据为一个1秒长度的显示波形，该波形会由手机的屏幕进行显示，这里可以在显示的时候将波形颜色设定为红色。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为绿光时，从第一绿光信号数据中，截取最新的、长度为显示时间长度的数据段，生成第一绿光显示数据；再对第一绿光显示数据，进行波形图像数据转换处理，生成第一绿光波形图像数据；然后将第一绿光波形图 像数据发送给显示屏14进行第一绿光波形显示处理。
例如，显示时间长度为1秒，光源信息为绿色，第一信号数据为第一绿光信号数据，则手机的控制单元会从第一绿光信号数据中截取最近一段时长为1秒的信号数据作为第一绿光显示数据，对应的第一绿光波形图像数据为一个1秒长度的显示波形，该波形会由手机的屏幕进行显示，这里可以在显示的时候将波形颜色设定为绿色。
在本实施例提供的另一个具体实现方式中，第一主控模块11具体用于当光源信息为红绿光时，从第二红光信号数据中，截取最新的、长度为显示时间长度的数据段，生成第二红光显示数据，再对第二红光显示数据，进行波形图像数据转换处理，生成第二红光波形图像数据；从第二绿光信号数据中，截取最新的、长度为显示时间长度的数据段，生成第二绿光显示数据，再对第二绿光显示数据，进行波形图像数据转换处理，生成第二绿光波形图像数据；然后将第二红光波形图像数据发送给显示屏14进行第二红光波形显示处理，将第二绿光波形图像数据发送给显示屏14进行第二绿光波形显示处理。
例如，显示时间长度为1秒，光源信息为绿色，第一信号数据包括了第二红光信号数据和第二绿光信号数据，则手机的控制单元会分别从第二红光信号数据和第二绿光信号数据中，各截取最近一段时长为1秒的信号数据作为第二红光显示数据和第二绿光显示数据，对应的，第二红光显示数据和第二绿光显示数据各自为一个1秒长度的显示波形，这两个波形会由手机的屏幕进行显示，为方便区分，这里可以在显示的时候，将第二红光显示数据的显示波形设定为红色，将第二绿光显示数据的显示波形设定为绿色。
第一主控模块11还用于将第一信号数据、第一设备令牌信息、第一设备类型信息、测试对象的第一年龄信息、第一性别信息、第一身高信息和第一体重信息，根据第一协议封装成第一数据包，并利用第一通讯模块15将第一数据包发送给云服务器2。
此处，第一设备令牌信息为对第一设备分配的合法设备令牌信息，第一设备类型信息为第一设备的具体设备类型信息，第一年龄信息、第一性别信息、第一身高信息和第一体重信息分别为测试对象的年龄、性别、身高和体重信息，第一协议为超文本传输协议(Hyper Text Transfer Protocol，HTTP)或超文本传输安全协议(Hyper Text Transfer Protocol over Secure Socket Layer，HTTPS)。
此处，第一通讯模块15具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网。
云服务器2包括第二主控模块21、合法性校验模块22、参数校验模块23、数据预处理模块24、人工智能血压预测模块25和第二通讯模块26。
例如，云服务器2具体为采用.NET内核叠加Tensorflow框架的云平台，则第二主控模块21为云平台的管理单元，合法性校验模块22为云平台的第一业务处理单元，参数校验模块23为云平台的第二业务处理单元，数据预处理模块24为云平台的数据处理单元，人工智能血压预测模块25为云平台的计算单元，第二通讯模块26具体为云平台的服务器通讯处理单元。
这里，.NET内核是一个通用的开放源代码开发平台，它由微软公司开发并提供，可以适配多种不同架构的处理器(诸如，美国英特尔公司x86与x64架构、英国ARM公司的ARM32与ARM64架构)，可以适配多种操作系统(诸如，视窗操作系统Windows、苹果电脑操作系统macOS和用户网络操作系统Linux)；TensorFlow是一个开源的人工智能模型学习框架，它由谷歌公司开发并提供，在图形分类、音频处理、推荐系统和自然语言处理等场景下有着丰富的应用，是目前主流的人工智能模型学习框架。
第二主控模块21用于将第一数据包按第一协议进行数据解析处理，得到第二信号数据、第二设备令牌信息、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息。
此处，第二信号数据等于第一信号数据；第二设备令牌信息等于第一 设备令牌信息；第二设备类型信息等于第一设备类型信息；第二年龄信息等于第一年龄信息；第二性别信息等于第一性别信息；第二身高信息等于第一身高信息；第二体重信息等于第一体重信息。
合法性校验模块22用于根据合法令牌列表，对第二设备令牌信息，进行合法性校验处理。
此处，合法令牌列表为一个存储了所有合法设备令牌信息的向量表。
在本实施例提供的另一个具体实现方式中，合法性校验模块22具体用于根据第二设备令牌信息，查询合法令牌列表，当第二设备令牌信息满足合法令牌列表时，合法性校验处理成功。
例如，第一业务处理单元从数据库中获得合法令牌列表，再对合法令牌列表中的所有合法设备令牌信息进行轮询，当发现第二设备令牌信息存在于合法令牌列表中时，合法性校验处理成功。
参数校验模块23用于当合法性校验处理成功时，对第二信号数据、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息，进行参数校验处理。
在本实施例提供的另一个具体实现方式中，参数校验模块23具体用于检查第二信号数据、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息是否全部不为空，当第二信号数据、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息全部不为空时，参数校验处理成功。
例如，第二业务处理单元在检查到第二信号数据、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息是否全部不为空，当第二信号数据、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息等均不为空时，参数完整性校验处理成功。
第二主控模块21还用于当参数完整性校验处理成功时，根据第二信号 数据，进行心率计算处理，生成心率数据。
例如，第二信号数据为长度为26秒的信号数据，管理单元将第二信号数据视为连续波形数据，依次提取该连续波形数据的波峰值的时间点作为信号时间点，将相邻信号时间点间的数据段作为间期数据段，统计26秒内的间期数据段的总数n，最终得到心率数据＝n*60/26(次/秒)。
数据预处理模块24用于根据预测模型标识信息，对第二信号数据、第二年龄信息、第二性别信息、第二身高信息和第二体重信息，进行血压预测模型输入数据准备处理，生成模型输入数据。
此处，预测模型标识信息至少包括第一卷积神经网络模型标识和第二卷积神经网络模型标识；数据预处理模块24包括多个子预处理模块。
在本实施例提供的另一个具体实现方式中，数据预处理模块24具体用于根据预测模型标识信息，选择对应的第一子预处理模块，对第二信号数据、第二年龄信息、第二性别信息、第二身高信息和第二体重信息，进行第一子血压预测模型输入数据准备处理，生成第一模型输入数据。
例如，预测模型标识信息为第一卷积神经网络模型标识，则数据处理单元选择具体为第一卷积神经网络模型预处理模块的第一子预处理模块，对第二信号数据、第二年龄信息、第二性别信息、第二身高信息和第二体重信息，进行第一卷积神经网络模型输入数据准备处理，生成具体为第一卷积神经网络模型输入数据的第一模型输入数据。
又例如，预测模型标识信息为第二卷积神经网络模型标识，则数据处理单元选择具体为第二卷积神经网络模型预处理模块的第一子预处理模块，对第二信号数据、第二年龄信息、第二性别信息、第二身高信息和第二体重信息，进行第二卷积神经网络模型输入数据准备处理，生成具体为第二卷积神经网络模型输入数据的第一模型输入数据。
在本实施例提供的另一个具体实现方式中，第一子预处理模块具体用于对第二信号数据进行消除基线漂移处理，生成第一过程信号数据，再对 第一过程信号数据进行降噪处理，生成第二过程信号数据，然后对第二过程信号数据进行标准采样和归一化处理，生成标准信号数据；再按第一子血压预测模型的输入数据格式要求，将第二信号数据、第二年龄信息、第二性别信息、第二身高信息、第二体重信息和标准信号数据封装成第一模型输入数据。
此处，第一子预处理模块中，消除基线漂移处理可以采用多项拟合滤波、中值滤波、无限脉冲响应(Infinite Impulse Response，IIR)滤波、快速傅里叶变换(fast Fourier transform，FFT)滤波或小波变换滤波等处理手段；降噪处理可以采用FFT滤波、带通滤波或带阻滤波等处理手段；归一化处理常见的是采用线性归一化方式；在做归一化处理之前需要对信号数据做一次标准采样，标准采样的采样频率与预处理子模块对应的血压预测模型的输入数据格式要求有关。
例如，预测模型标识信息为第一卷积神经网络模型标识，第一子预处理模块为第一卷积神经网络模型预处理模块，则第一子预处理模块采用多项拟合滤波、中值滤波、无限脉冲响应滤波、快速傅里叶变换滤波或小波变换滤波等处理手段对第二信号数据进行消除基线漂移处理生成第一过程信号数据；再采用快速傅里叶变换滤波、带通滤波或带阻滤波等处理手段对第一过程信号数据进行降噪处理生成第二过程信号数据；再采用与第一卷积神经网络模型对应的采样频率对第二过程信号数据进行标准采样，然后通过线性归一化方式对采样后数据进行归一化处理生成标准信号数据；最后按第一卷积神经网络模型的输入数据格式要求，将第二信号数据、第二年龄信息、第二性别信息、第二身高信息、第二体重信息和标准信号数据封装成具体为第一卷积神经网络模型输入数据的第一模型输入数据。
又例如，预测模型标识信息为第二卷积神经网络模型标识，第一子预处理模块为第二卷积神经网络模型预处理模块，则第一子预处理模块采用多项拟合滤波、中值滤波、无限脉冲响应滤波、快速傅里叶变换滤波或小波变换 滤波等处理手段对第二信号数据进行消除基线漂移处理生成第一过程信号数据；再采用快速傅里叶变换滤波、带通滤波或带阻滤波等处理手段对第一过程信号数据进行降噪处理生成第二过程信号数据；再采用与第二卷积神经网络模型对应的采样频率对第二过程信号数据进行标准采样，然后通过线性归一化方式对采样后数据进行归一化处理生成标准信号数据；最后按第二卷积神经网络模型的输入数据格式要求，将第二信号数据、第二年龄信息、第二性别信息、第二身高信息、第二体重信息和标准信号数据封装成具体为第二卷积神经网络模型输入数据的第一模型输入数据。
人工智能血压预测模块25用于根据预测模型标识信息，对模型输入数据，进行血压预测运算处理，生成舒张压数据和收缩压数据。
此处，人工智能血压预测模块25包括多个子血压预测模型。
这里，人工智能血压预测模块25至少提供2种血压预测模型：第一卷积神经网络模型和第二卷积神经网络模型：
(一)第一卷积神经网络模型：包括多层卷积神经网络层和全连接层，每层卷积神经网络层包括1层卷积层和1层池化层；卷积层负责对模型输入数据进行血压特征提取计算，池化层则是对卷积层的提取结果进行降采样，每一层卷积神经网络层的输出结果作为下一层卷积神经网络层的输入，最后的卷积神经网络层计算结果输入到全连接层进行回归计算，得到收缩压数据和舒张压数据；
(二)第二卷积神经网络模型：包括二维卷积层、最大池化层、批量均一化层、激活层、相加层、全局平均池化层、随机丢弃层和全连接层，二维卷积层可以包含多个子卷积层，负责对模型输入数据进行多次卷积计算，二维卷积层输出的卷积结果包含多个一维向量，最大池化层在每个一维向量里以取最大值的方式对卷积结果进行采样起到降低数据量的作用，批量均一化层是对由最大池化层的输出结果进行数据均一化处理，激活层采用非线性激活函数对批量均一化层的输出结果进行神经网络连接，相加层对 激活层输出结果进行加权相加计算，全局平均池化层对相加层输出结果进行全数据加权平均计算，随机丢弃层按随机性将全局平均池化层的输出结果进行裁剪，最终使用全连接层对裁剪后的随机丢弃层输出结果进行二分类回归计算输出舒张压数据和收缩压数据。
在本实施例提供的另一个具体实现方式中，人工智能血压预测模块25具体用于根据预测模型标识信息，选择对应的第一子血压预测模型，对第一模型输入数据，进行第一血压预测运算处理，生成舒张压数据和收缩压数据。
例如，预测模型标识信息为第一卷积神经网络模型标识，第一模型输入数据为第一卷积神经网络模型输入数据，计算单元选择具体为第一卷积神经网络模型的第一子血压预测模型，对第一卷积神经网络模型输入数据，进行第一卷积神经网络血压预测运算处理，生成舒张压数据和收缩压数据。
又例如，预测模型标识信息为第二卷积神经网络模型标识，第一模型输入数据为第二卷积神经网络模型输入数据，计算单元选择具体为第二卷积神经网络模型的第一子血压预测模型，对第二卷积神经网络模型输入数据，进行第二卷积神经网络血压预测运算处理，生成舒张压数据和收缩压数据。
第二主控模块21还用于设置状态码数据为正常状态码信息，然后根据心率数据、舒张压数据和收缩压数据组成返回数据，再将返回数据和状态码数据，根据第一协议封装成第二数据包，并利用第二通讯模块26将第二数据包发送给第一设备。
在本实施例提供的另一个具体实现方式中，第二通讯模块26具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网。
例如，心率数据为76次/分，舒张压数据为85毫米汞柱，收缩压数据为112毫米汞柱，则云平台的管理单元将状态码数据设置为正常状态码信息，设置返回数据由心率数据+舒张压数据+收缩压数据组成，并按HTTP或HTTPS数据打包格式对返回数据和状态码数据进行打包，生成第二数据包；并利用云平台的通讯处理单元向手机发送第二数据包。
第一主控模块11还用于将第二数据包按第一协议进行数据解析处理，得到返回数据和状态码数据；当状态码数据为正常状态码信息时、从返回数据中获取心率数据、舒张压数据和收缩压数据；再将心率数据、舒张压数据和收缩压数据发送给显示屏14进行心率及血压数据显示处理。
例如，手机的控制单元在解析完第二数据包之后，得到心率数据为76次/分，舒张压数据为85毫米汞柱，收缩压数据为112毫米汞柱；手机的屏幕在接收到控制单元发送的数据之后，对以下信息进行显示：“心率：76次/分”、“舒张压：85毫米汞柱”和“收缩压：112毫米汞柱”。
本发明实施例提供的一种基于视频数据进行血压预测的系统，包括第一设备和云服务器。其中，第一设备使用第一摄像头对测试对象的表皮区域进行视频拍摄得到第一视频数据，再对第一视频数据进行远程光体积描计信号数据转换处理得到第一信号数据；云服务器则对第一信号数据进行心率分析得到心率数据，并使用人工智能血压预测模型，对结合了测试对象个性数据(年龄、性别、身高、体重等信息)和第一信号数据的模型输入数据，进行血压分析得到血压数据(收缩压、舒张压)；最后，云服务器将分析数据(心率数据、收缩压数据、舒张压数据)返回到第一设备进行显示。使用该系统，测试对象无需佩戴任何特定的采集装置，通过第一设备可随时启动对测试对象的实时检测或持续监测，降低了对测试对象进行血压实时检测或持续监测的难度，丰富了光体积描计法在监护领域的应用场景。
专业人员应该还可以进一步意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、计算机软件或者二者的结合来实现，为了清楚地说明硬件和软件的可互换性，在上述说明中已经按照功能一般性地描述了各示例的组成及步骤。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。
结合本文中所公开的实施例描述的方法或算法的步骤可以用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。
以上所述的具体实施方式，对本发明的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本发明的具体实施方式而已，并不用于限定本发明的保护范围，凡在本发明的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (10)

  1. 一种基于视频数据进行血压预测的系统，其特征在于，所述基于视频数据进行血压预测的系统包括：第一设备和云服务器；
    所述第一设备包括第一主控模块、第一摄像头、灯光模块、显示屏和第一通讯模块；
    所述第一主控模块用于调用所述第一摄像头和所述灯光模块，对测试对象的表皮区域进行第一持续时间的拍摄处理，从而生成第一视频数据；
    所述显示屏用于接收所述主控模块发送的所述第一视频数据，进行播放处理；
    所述第一主控模块还用于根据光源信息，对所述第一视频数据进行光源通道数据提取处理，生成第一通道数据；然后对所述第一通道数据，进行远程光体积描计信号数据转换处理，生成第一信号数据；再根据显示时间长度将所述第一信号数据发送给所述显示屏进行信号波形显示处理；
    所述第一主控模块还用于将所述第一信号数据、第一设备令牌信息、第一设备类型信息、所述测试对象的第一年龄信息、第一性别信息、第一身高信息和第一体重信息，根据第一协议封装成第一数据包，并利用所述第一通讯模块将所述第一数据包发送给所述云服务器；
    所述云服务器包括第二主控模块、合法性校验模块、参数校验模块、数据预处理模块、人工智能血压预测模块和第二通讯模块；
    所述第二主控模块用于将所述第一数据包按所述第一协议进行数据解析处理，得到第二信号数据、第二设备令牌信息、第二设备类型信息、第二年龄信息、第二性别信息、第二身高信息和第二体重信息；
    所述合法性校验模块用于根据合法令牌列表，对所述第二设备令牌信息，进行合法性校验处理；
    所述参数校验模块用于当所述合法性校验处理成功时，对所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、 所述第二身高信息和所述第二体重信息，进行参数校验处理；
    所述第二主控模块还用于当所述参数完整性校验处理成功时，根据所述第二信号数据，进行心率计算处理，生成心率数据；
    所述数据预处理模块用于根据预测模型标识信息，对所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息，进行血压预测模型输入数据准备处理，生成模型输入数据；
    所述人工智能血压预测模块用于根据所述预测模型标识信息，对所述模型输入数据，进行血压预测运算处理，生成舒张压数据和收缩压数据；
    所述第二主控模块还用于设置状态码数据为正常状态码信息，然后根据所述心率数据、所述舒张压数据和所述收缩压数据组成返回数据，再将所述返回数据和所述状态码数据，根据所述第一协议封装成第二数据包，并利用所述第二通讯模块将所述第二数据包发送给所述第一设备；
    所述第一主控模块还用于将所述第二数据包按所述第一协议进行数据解析处理，得到所述返回数据和所述状态码数据；当所述状态码数据为所述正常状态码信息时、从所述返回数据中获取所述心率数据、所述舒张压数据和所述收缩压数据；再将所述心率数据、所述舒张压数据和所述收缩压数据发送给所述显示屏进行心率及血压数据显示处理。
  2. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一主控模块具体用于调用所述灯光模块对所述测试对象的表皮区域进行照射，并在所述第一摄像头的镜头覆盖了所述表皮区域之后，利用所述第一摄像头对所述表皮区域进行所述第一持续时间的拍摄处理，从而生成所述第一视频数据。
  3. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一主控模块具体用于当所述光源信息为红光时，对所述第一视频 数据进行红光通道数据提取处理，生成所述第一通道数据；当所述光源信息为绿光时，对所述第一视频数据进行绿光通道数据提取处理，生成所述第一通道数据；当所述光源信息为红绿光时，对所述第一视频数据进行红光通道数据提取处理生成第一红光通道数据，对所述第一视频数据进行绿光通道数据提取处理生成第一绿光通道数据，并将所述第一红光通道数据和所述第一绿光通道数据封装成所述第一通道数据。
  4. 根据权利要求3所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一主控模块具体用于当所述光源信息为所述红光时，对所述第一视频数据进行帧图像提取处理，得到多个第一帧图像数据；在每个第一帧图像数据中，对像素值满足红光像素阈值范围的第一红色像素点的数量进行统计，生成第一总数，并对所有所述第一红色像素点进行像素值总和计算，生成第一像素值总和，再将所述第一像素值总和与所述第一总数的比值，做为与所述每个第一帧图像数据对应的第一帧红光通道数据；然后对所有所述第一帧红光通道数据，按时间先后顺序排列，生成所述第一通道数据；
    所述第一主控模块具体用于当所述光源信息为所述绿光时，对所述第一视频数据进行帧图像提取处理，得到多个第二帧图像数据；在每个第二帧图像数据中，对像素值满足绿光像素阈值范围的第一绿色像素点的数量进行统计，生成第二总数，并对所有所述第一绿色像素点进行像素值总和计算，生成第二像素值总和，再将所述第二像素值总和与所述第二总数的比值，做为与所述每个第二帧图像数据对应的第一帧绿光通道数据；然后对所有所述第一帧绿光通道数据，按时间先后顺序排列，生成所述第一通道数据；
    所述第一主控模块具体用于当所述光源信息为所述红绿光时，对所述第一视频数据进行视频帧图像提取处理，得到多个第三帧图像数据；在每个第三帧图像数据中，对像素值满足所述红光像素阈值范围的第二红色像素点的数量进行统计，生成第三总数，并对所有所述第二红色像素点进行像素值总 和计算，生成第三像素值总和，再将所述第三像素值总和与所述第三总数的比值，做为与所述每个第三帧图像数据对应的第二帧红光通道数据，然后对所有所述第二帧红光通道数据，按时间先后顺序排列，生成所述第一红光通道数据；并在所述每个第三帧图像数据中，对像素值满足所述绿光像素阈值范围的第二绿色像素点的数量进行统计，生成第四总数，并对所有所述第二绿色像素点进行像素值总和计算，生成第四像素值总和，再将所述第四像素值总和与所述第四总数的比值，做为与所述每个第三帧图像数据对应的第二帧绿光通道数据，然后对所有所述第二帧绿光通道数据，按时间先后顺序排列，生成所述第一绿光通道数据；再对所述第一红光通道数据和所述第一绿光通道数据，进行多通道数据封装处理，生成所述第一通道数据。
  5. 根据权利要求4所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一主控模块具体用于当所述光源信息为所述红光时，对所述第一通道数据进行远程光体积描计信号带通滤波处理，生成第一红光滤波数据，再对所述第一红光滤波数据进行远程光体积描计信号降噪处理，生成第一红光信号数据；
    所述第一主控模块具体用于当所述光源信息为所述绿光时，对所述第一通道数据进行远程光体积描计信号带通滤波处理，生成第一绿光滤波数据，再对所述第一绿光滤波数据进行远程光体积描计信号降噪处理，生成第一绿光信号数据；
    所述第一主控模块具体用于当所述光源信息为所述红绿光时，对所述第一通道数据进行红光通道数据提取处理，生成第二红光通道数据；并对所述第一通道数据进行绿光通道数据提取处理，生成第二绿光通道数据；再对所述第二红光通道数据和所述第二绿光通道数据，分别进行远程光体积描计信号带通滤波处理，生成第二红光滤波数据和第二绿光滤波数据；然后对所述第二红光滤波数据和所述第二绿光滤波数据，分别进行远程光体积描计信号 降噪处理，生成第二红光信号数据和第二绿光信号数据。
  6. 根据权利要求5所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一主控模块具体用于当所述光源信息为所述红光时，从所述第一红光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第一红光显示数据；再对所述第一红光显示数据，进行波形图像数据转换处理，生成第一红光波形图像数据；然后将所述第一红光波形图像数据发送给所述显示屏进行第一红光波形显示处理；
    所述第一主控模块具体用于当所述光源信息为所述绿光时，从所述第一绿光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第一绿光显示数据；再对所述第一绿光显示数据，进行波形图像数据转换处理，生成第一绿光波形图像数据；然后将所述第一绿光波形图像数据发送给所述显示屏进行第一绿光波形显示处理；
    所述第一主控模块具体用于当所述光源信息为所述红绿光时，从所述第二红光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第二红光显示数据，再对所述第二红光显示数据，进行波形图像数据转换处理，生成第二红光波形图像数据；从所述第二绿光信号数据中，截取最新的、长度为所述显示时间长度的数据段，生成第二绿光显示数据，再对所述第二绿光显示数据，进行波形图像数据转换处理，生成第二绿光波形图像数据；然后将所述第二红光波形图像数据发送给所述显示屏进行第二红光波形显示处理，将所述第二绿光波形图像数据发送给所述显示屏进行第二绿光波形显示处理。
  7. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述第一协议包括超文本传输协议HTTP和超文本传输安全协议HTTPS；
    所述第二信号数据等于所述第一信号数据；所述第二设备令牌信息等于所述第一设备令牌信息；所述第二设备类型信息等于所述第一设备类型信息；所述第二年龄信息等于所述第一年龄信息；所述第二性别信息等于所述第一性别信息；所述第二身高信息等于所述第一身高信息；所述第二体重信息等于所述第一体重信息；
    所述第一通讯模块具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网；
    所述第二通讯模块具体用于通过移动通信网络、无线局域网络或有线局域网络接入互联网。
  8. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述合法性校验模块具体用于根据所述第二设备令牌信息，查询所述合法令牌列表，当所述第二设备令牌信息满足所述合法令牌列表时，所述合法性校验处理成功。
  9. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述参数校验模块具体用于检查所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息是否全部不为空，当所述第二信号数据、所述第二设备类型信息、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息全部不为空时，所述参数校验处理成功。
  10. 根据权利要求1所述的基于视频数据进行血压预测的系统，其特征在于，
    所述数据预处理模块包括多个子预处理模块；所述人工智能血压预测模块包括多个子血压预测模型；
    所述数据预处理模块具体用于根据预测模型标识信息，选择对应的第一 子预处理模块，对所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息和所述第二体重信息，进行第一子血压预测模型输入数据准备处理，生成第一模型输入数据；
    所述第一子预处理模块具体用于对所述第二信号数据进行消除基线漂移处理，生成第一过程信号数据，再对所述第一过程信号数据进行降噪处理，生成第二过程信号数据，然后对所述第二过程信号数据进行标准采样和归一化处理，生成标准信号数据；再按第一子血压预测模型的输入数据格式要求，将所述第二信号数据、所述第二年龄信息、所述第二性别信息、所述第二身高信息、所述第二体重信息和所述标准信号数据封装成所述第一模型输入数据；
    所述人工智能血压预测模块具体用于根据所述预测模型标识信息，选择对应的第一子血压预测模型，对所述第一模型输入数据，进行第一血压预测运算处理，生成所述舒张压数据和所述收缩压数据。
PCT/CN2021/088017 2020-10-12 2021-04-19 一种基于视频数据进行血压预测的系统 WO2022077887A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US18/248,713 US20230389808A1 (en) 2020-10-12 2021-04-19 Video data-based system for blood pressure prediction
EP21878941.0A EP4226855A1 (en) 2020-10-12 2021-04-19 Video data-based system for blood pressure prediction

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011086786.5A CN112315437A (zh) 2020-10-12 2020-10-12 一种基于视频数据进行血压预测的系统
CN202011086786.5 2020-10-12

Publications (1)

Publication Number Publication Date
WO2022077887A1 true WO2022077887A1 (zh) 2022-04-21

Family

ID=74314856

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/088017 WO2022077887A1 (zh) 2020-10-12 2021-04-19 一种基于视频数据进行血压预测的系统

Country Status (4)

Country Link
US (1) US20230389808A1 (zh)
EP (1) EP4226855A1 (zh)
CN (1) CN112315437A (zh)
WO (1) WO2022077887A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112315437A (zh) * 2020-10-12 2021-02-05 乐普（北京）医疗器械股份有限公司 一种基于视频数据进行血压预测的系统
CN113456042A (zh) * 2021-06-30 2021-10-01 浙江师范大学 一种基于3d cnn的无接触面部血压测量方法
CN114397334B (zh) * 2021-12-28 2024-05-07 乐普（北京）医疗器械股份有限公司 一种无创血糖分析系统

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106343989A (zh) * 2016-10-18 2017-01-25 北京博瑞彤芸文化传播股份有限公司 基于图像处理的血压监测方法
CN107049281A (zh) * 2017-05-25 2017-08-18 姚丽鹏 一种用于生理信息监测的智能眼镜
US20180295134A1 (en) * 2017-04-07 2018-10-11 Citrix Systems, Inc. Systems and methods for securely and transparently proxying saas applications through a cloud-hosted or on-premise network gateway for enhanced security and visibility
US20180303351A1 (en) * 2017-04-20 2018-10-25 General Electric Company Systems and methods for optimizing photoplethysmograph data
CN109009034A (zh) * 2018-07-10 2018-12-18 京东方科技集团股份有限公司 血压测量方法、终端及存储介质
CN110251105A (zh) * 2019-06-12 2019-09-20 广州视源电子科技股份有限公司 一种无创血压测量方法、装置、设备及系统
CN111179454A (zh) * 2019-12-10 2020-05-19 深圳技术大学 签到和生理参数检测系统及其控制方法
CN111297339A (zh) * 2020-02-21 2020-06-19 乐普（北京）医疗器械股份有限公司 一种光体积变化描记图法信号的生成方法和装置
CN111297347A (zh) * 2020-02-21 2020-06-19 乐普（北京）医疗器械股份有限公司 一种生成光体积变化描记图法信号的方法和装置
CN111358455A (zh) * 2020-03-17 2020-07-03 乐普（北京）医疗器械股份有限公司 一种多数据源的血压预测方法和装置
CN112315437A (zh) * 2020-10-12 2021-02-05 乐普（北京）医疗器械股份有限公司 一种基于视频数据进行血压预测的系统

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106343989A (zh) * 2016-10-18 2017-01-25 北京博瑞彤芸文化传播股份有限公司 基于图像处理的血压监测方法
US20180295134A1 (en) * 2017-04-07 2018-10-11 Citrix Systems, Inc. Systems and methods for securely and transparently proxying saas applications through a cloud-hosted or on-premise network gateway for enhanced security and visibility
US20180303351A1 (en) * 2017-04-20 2018-10-25 General Electric Company Systems and methods for optimizing photoplethysmograph data
CN107049281A (zh) * 2017-05-25 2017-08-18 姚丽鹏 一种用于生理信息监测的智能眼镜
CN109009034A (zh) * 2018-07-10 2018-12-18 京东方科技集团股份有限公司 血压测量方法、终端及存储介质
CN110251105A (zh) * 2019-06-12 2019-09-20 广州视源电子科技股份有限公司 一种无创血压测量方法、装置、设备及系统
CN111179454A (zh) * 2019-12-10 2020-05-19 深圳技术大学 签到和生理参数检测系统及其控制方法
CN111297339A (zh) * 2020-02-21 2020-06-19 乐普（北京）医疗器械股份有限公司 一种光体积变化描记图法信号的生成方法和装置
CN111297347A (zh) * 2020-02-21 2020-06-19 乐普（北京）医疗器械股份有限公司 一种生成光体积变化描记图法信号的方法和装置
CN111358455A (zh) * 2020-03-17 2020-07-03 乐普（北京）医疗器械股份有限公司 一种多数据源的血压预测方法和装置
CN112315437A (zh) * 2020-10-12 2021-02-05 乐普（北京）医疗器械股份有限公司 一种基于视频数据进行血压预测的系统

Also Published As

Publication number Publication date
CN112315437A (zh) 2021-02-05
EP4226855A1 (en) 2023-08-16
US20230389808A1 (en) 2023-12-07

Similar Documents

Publication Publication Date Title
WO2022077887A1 (zh) 一种基于视频数据进行血压预测的系统
Wang et al. A comparative survey of methods for remote heart rate detection from frontal face videos
US9020185B2 (en) Systems and methods for non-contact heart rate sensing
EP2748762B1 (en) Distortion reduced signal detection
US10004410B2 (en) System and methods for measuring physiological parameters
CN105813564A (zh) 用于确定对象的生命体征的设备和方法
Casado et al. Face2PPG: An unsupervised pipeline for blood volume pulse extraction from faces
Wang et al. Detail-preserving pulse wave extraction from facial videos using consumer-level camera
US20220133161A1 (en) Systems and methods for acquiring ppg signals for measuring blood pressure
US20200105400A1 (en) Fully Automated Non-Contact Remote Biometric and Health Sensing Systems, Architectures, and Methods
US20230233091A1 (en) Systems and Methods for Measuring Vital Signs Using Multimodal Health Sensing Platforms
WO2021164350A1 (zh) 一种生成光体积变化描记图法信号的方法和装置
CN114402359A (zh) 用于检测人的合成视频的系统和方法
Bousefsaf et al. iPPG 2 cPPG: reconstructing contact from imaging photoplethysmographic signals using U-Net architectures
Karthick et al. Analysis of vital signs using remote photoplethysmography (RPPG)
CN116109818A (zh) 一种基于面部视频的中医脉候判别系统及方法和装置
Abdulrahaman Two-stage motion artifact reduction algorithm for rPPG signals obtained from facial video recordings
JP2023505111A (ja) 光学データからの生理学的測定のためのシステムおよび方法
Ayesha ReViSe: An end-to-end framework for remote measurement of vital signs
Ayesha et al. A web application for experimenting and validating remote measurement of vital signs
WO2017051415A1 (en) A system and method for remotely obtaining physiological parameter of a subject
Lee et al. Review on remote heart rate measurements using photoplethysmography
Jia et al. CEBPM: A Cloud-Edge Collaborative Noncontact Blood Pressure Estimation Model
US20240161498A1 (en) Non-contrastive unsupervised learning of physiological signals from video
El Khadiri et al. An efficient hybrid algorithm for non-contact physiological sign monitoring using plethysmography wave analysis

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21878941

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2021878941

Country of ref document: EP

Effective date: 20230512