WO2017084546A1 - 基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备 - Google Patents

基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备 Download PDF

Info

Publication number
WO2017084546A1
WO2017084546A1 PCT/CN2016/105720 CN2016105720W WO2017084546A1 WO 2017084546 A1 WO2017084546 A1 WO 2017084546A1 CN 2016105720 W CN2016105720 W CN 2016105720W WO 2017084546 A1 WO2017084546 A1 WO 2017084546A1
Authority
WO
WIPO (PCT)
Prior art keywords
feature vector
ecg signal
wave
original
user
Prior art date
Application number
PCT/CN2016/105720
Other languages
English (en)
French (fr)
Inventor
赵亚军
苏吉祥
王飞
陈婷
毛红达
Original Assignee
安徽华米信息科技有限公司
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
Priority claimed from CN201510795936.2A external-priority patent/CN105286853B/zh
Priority claimed from CN201510796544.8A external-priority patent/CN105468951B/zh
Application filed by 安徽华米信息科技有限公司 filed Critical 安徽华米信息科技有限公司
Priority to US15/584,911 priority Critical patent/US10163528B2/en
Publication of WO2017084546A1 publication Critical patent/WO2017084546A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
    • 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
    • A61B5/02438—Detecting, measuring or recording pulse rate or heart rate with portable devices, e.g. worn by the patient
    • 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
    • A61B5/0245—Detecting, measuring or recording pulse rate or heart rate by using sensing means generating electric signals, i.e. ECG signals
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/321—Accessories or supplementary instruments therefor, e.g. cord hangers
    • A61B5/322—Physical templates or devices for measuring ECG waveforms, e.g. electrocardiograph rulers or calipers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A61B5/349—Detecting specific parameters of the electrocardiograph cycle
    • 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
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/726—Details of waveform analysis characterised by using transforms using Wavelet transforms
    • 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
    • A61B5/7271—Specific aspects of physiological measurement analysis
    • A61B5/7282—Event detection, e.g. detecting unique waveforms indicative of a medical condition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/147—Discrete orthonormal transforms, e.g. discrete cosine transform, discrete sine transform, and variations therefrom, e.g. modified discrete cosine transform, integer transforms approximating the discrete cosine transform
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/148—Wavelet transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/15—Correlation function computation including computation of convolution operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
    • G06F21/31—User authentication
    • G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
    • G16H40/63—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for local operation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
    • G16Z99/00—Subject matter not provided for in other main groups of this subclass
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/30—Circuit design
    • G06F30/32—Circuit design at the digital level
    • G06F30/33—Design verification, e.g. functional simulation or model checking
    • G06F30/3308—Design verification, e.g. functional simulation or model checking using simulation
    • G06F30/331—Design verification, e.g. functional simulation or model checking using simulation with hardware acceleration, e.g. by using field programmable gate array [FPGA] or emulation

Definitions

  • This application relates to the field of wearable device technology.
  • a method for determining user attention information based on a wearable device which is applied to the wearable device, and includes:
  • Determining a feature vector of the original ECG signal the feature vector including time domain feature data corresponding to the original ECG signal and frequency domain feature data of the original ECG signal;
  • the user's attention information is determined based on the similarity between the feature vector and the reference feature vector representing the attention information.
  • a wearable device comprising:
  • An electrocardiogram sensor for collecting a user's original ECG signal
  • An FPGA system configured to determine a user's attention information based on the original ECG signal, including:
  • a feature vector extraction module configured to determine a feature vector of the original ECG signal, the feature vector including time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal;
  • the result discriminating module is configured to determine the user's attention information based on the similarity between the feature vector and the reference feature vector representing the attention information.
  • a user attention information determining apparatus based on a wearable device which is applied to the wearable device, and includes:
  • a memory for storing machine executable instructions
  • the processor is caused to:
  • Determining a feature vector of the original ECG signal the feature vector including time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal;
  • the user's attention information is determined based on the similarity between the feature vector and the reference feature vector representing the attention information.
  • the present application can collect the original ECG signal of the user through the ECG sensor, and determine the feature vector of the original ECG signal, which can be based on the similarity between the feature vector and the reference feature vector representing the attention information. Determine the user's attention information. Since the feature vector can include time domain feature data and frequency domain feature data of the original ECG signal, the matrix model used in the similarity algorithm can be obtained by a machine learning method, which can make the user's attention information determined by the ECG accurate. Higher degrees.
  • FIG. 1 is a flow chart showing a method for determining user interest information based on a wearable device according to an exemplary embodiment of the present application.
  • FIG. 2A shows a schematic flowchart of a wearable device-based disease detection method according to an exemplary embodiment of the present application
  • FIG. 2B shows a schematic diagram of an original electrocardiographic signal in accordance with an exemplary embodiment of the present application
  • FIG. 3A illustrates a flow diagram of determining frequency domain feature data of an original electrocardiographic signal, in accordance with an exemplary embodiment of the present application
  • FIG. 3B illustrates a schematic diagram of an ECG signal after noise filtering by wavelet transform, in accordance with an exemplary embodiment of the present application
  • FIG. 4A illustrates a flow diagram of determining time domain feature data of an original electrocardiographic signal, in accordance with an exemplary embodiment of the present application
  • 4B is a schematic diagram showing timing and amplitude characteristics of an electrocardiographic signal
  • 4C is a schematic diagram showing detection of R wave peaks by a hardware circuit
  • Figure 4D shows a circuit diagram for detecting a dynamic threshold
  • FIG. 5 is a flow chart showing a wearable device-based disease detection method according to still another exemplary embodiment of the present invention.
  • FIG. 6 shows a schematic flowchart of a wearable device-based identification method according to an exemplary embodiment of the present application
  • Figure 7 shows a schematic diagram of the ECG signals of two different users
  • FIG. 8 is a block diagram showing the structure of a wearable device according to an exemplary embodiment of the present application.
  • FIG. 9 is a block diagram showing a hardware structure of a device for determining user interest information based on a wearable device according to an exemplary embodiment of the present application.
  • FIG. 10 illustrates a functional block diagram of a wearable device-based user attention information determination logic, in accordance with an exemplary embodiment of the present application
  • FIG. 11 illustrates a functional block diagram of a wearable device-based user attention information determination logic in accordance with yet another exemplary embodiment of the present application.
  • FIG. 1 is a flow chart showing a method for determining user interest information based on a wearable device according to an exemplary embodiment of the present application.
  • the method can include:
  • Step 101 Acquire an original ECG signal of the user through the ECG sensor integrated by the wearable device.
  • Step 102 Determine a feature vector of the original ECG signal, where the feature vector may include time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal.
  • Step 103 Determine the user's attention information based on the similarity between the feature vector and the reference feature vector representing the attention information.
  • the attention information may be health information, including whether the user has a certain disease.
  • the attention information may also be identity information, including whether it is a legitimate user or the like.
  • the human body's electrocardiogram can be determined by the individual's cardiac structure, which is universal, unique, easy to collect, and permanent. Moreover, the ECG also has advantages such as being only in the living body, being difficult to be copied, and not easily lost. Therefore, the ECG can be collected by the wearable device, and the cardiovascular disease of the user can be detected based on the ECG, so that the user can find the abnormality of the body in time, and ensure that the user can Get medical treatment in time.
  • the application can collect the original ECG signal of the user through the ECG sensor integrated in the wearable device, determine the feature vector of the original ECG signal, and determine the originality according to the similarity between the feature vector and the reference feature vector of the disease ECG signal.
  • the type of disease corresponding to the ECG signal Since the feature vector includes time domain feature data and frequency domain feature data of the original ECG signal, the user can perform disease detection through the ECG-based feature vector, which can greatly improve the accuracy of the user's disease detection, thereby implementing the disease based on the wearable device. Identify as early as possible.
  • FIG. 2A shows a flow diagram of a wearable device-based disease detection method in accordance with an exemplary embodiment of the present application.
  • 2B shows a schematic diagram of an original electrocardiographic signal in accordance with an exemplary embodiment of the present application.
  • the wearable device-based disease detection method can be applied to a wearable device, such as a smart bracelet, and the like, and the ECG sensor can be disposed on the wearable device.
  • the disease detection method may include the following steps:
  • step 201 the user's original ECG signal is collected by the ECG sensor.
  • Step 202 determining a feature vector of the original ECG signal.
  • the feature vector may include time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal.
  • Step 203 Determine a disease type corresponding to the original ECG signal based on the similarity between the feature vector and the reference feature vector of the disease ECG signal.
  • the original ECG signal collected by the ECG sensor generally has strong noise, and the noise may change over time, but the QRS of the original ECG signal collected by the same user at different times.
  • the wave group, the P wave, and the T wave are basically the same.
  • the frequency domain feature data of the original ECG signal may include a wavelet coefficient corresponding to the original ECG signal, a discrete cosine transform coefficient, a Fourier transform coefficient, a HHT (Hilbert-Hwang) transform coefficient, and the like.
  • the present application does not limit the specific manner of transformation.
  • the existing user's ECG data can be trained to obtain reference feature vectors for the ECG signals of various diseases.
  • the reference feature vector of the disease ECG signal can be completed offline, and the reference feature vector of the disease ECG signal obtained by the training can be directly used when implementing the disease detection.
  • the matrix model used in the nearest neighbor algorithm's similarity algorithm can be obtained by machine learning. The machine learning goal is based on the accuracy of the similarity algorithm as high as possible, thus ensuring the accuracy of disease detection.
  • the embodiment of the present application collects the original ECG signal of the user through the ECG sensor to determine the original heart. a feature vector of the electrical signal, and calculating a similarity between the feature vector and a reference feature vector of each disease ECG signal according to a similarity algorithm such as a nearest neighbor algorithm, so that the original heart can be determined according to the similarity
  • a similarity algorithm such as a nearest neighbor algorithm
  • the type of disease corresponding to the electrical signal Since the feature vector can include time domain feature data and frequency domain feature data of the original ECG signal, the matrix model used in the similarity algorithm can be obtained by a machine learning method, so that the accuracy of detecting the disease by the ECG is better. high.
  • FIG. 3A illustrates a flow diagram of determining frequency domain feature data of an original electrocardiographic signal, in accordance with an exemplary embodiment of the present application.
  • FIG. 3B illustrates a schematic diagram of an ECG signal after noise filtering by wavelet transform, in accordance with an exemplary embodiment of the present application.
  • determining frequency domain characteristic data of the original ECG signal may include the following steps:
  • Step 301 Perform wavelet transformation on the original ECG signal to obtain a wavelet coefficient of the original ECG signal.
  • Step 302 Determine the wavelet coefficient as a frequency domain characteristic data of the original ECG signal.
  • Step 303 performing autocorrelation and discrete cosine transform on the wavelet transformed ECG signal to obtain discrete cosine transform coefficients after autocorrelation and discrete cosine transform.
  • Step 304 Determine the discrete cosine transform coefficient as another frequency domain feature data of the original ECG signal.
  • step 301 signals of different frequencies in the original ECG signal can be decomposed by wavelet transform. Since the high-frequency noise of the original ECG signal is mainly reflected on the low scale, the low-frequency noise of the original ECG signal is mainly reflected on the high scale, and the analysis of the original ECG signal by the intermediate scale after the wavelet transform can be effectively useful.
  • the signal is separated from the interference signal.
  • the original ECG signal may be wavelet-decomposed by a set of high-pass and low-pass filters predetermined to obtain a wavelet coefficient corresponding to the original ECG signal.
  • the wavelet coefficients may include wavelet coefficients on a multi-level scale.
  • the wavelet transform of the original ECG signal can be implemented by means of shifting and adding the hardware platform of the FPGA, and the logic of the shifting and adding operations is simple and easy to implement. Referring to FIG. 3B, after the wavelet transform, the noise of the original ECG signal can be effectively removed, and the obtained ECG signal is more regular.
  • the filtered electrocardiographic signal may be subjected to an autocorrelation operation, and the autocorrelation operation may eliminate the signal portion of the electrocardiographic signal that is unrelated to the identification. Then, the electrocardiographic signal after the autocorrelation operation can be subjected to discrete cosine transform to obtain discrete cosine transform coefficients.
  • the filtered ECG signal may also be subjected to Fourier transform, HHT transform, or the like, and the transformed Fourier transform coefficient or HHT transform coefficient is used as a frequency domain characteristic data of the ECG signal.
  • the noise of the original ECG signal can be filtered by the wavelet transform, so that the ECG signal is more regular, thereby ensuring that the original wavelet signal and the discrete cosine transform coefficient of the original ECG signal represent the original heart more accurately.
  • the characteristics of electrical signals in the frequency domain can be filtered by the wavelet transform, so that the ECG signal is more regular, thereby ensuring that the original wavelet signal and the discrete cosine transform coefficient of the original ECG signal represent the original heart more accurately.
  • FIG. 4A illustrates a flow diagram of determining time domain feature data of an original electrocardiographic signal, in accordance with an exemplary embodiment of the present application.
  • Figure 4B shows a schematic of the timing and amplitude characteristics of the ECG signal.
  • Figure 4C shows a schematic diagram of detecting R wave peaks using a hardware circuit.
  • Figure 4D shows a circuit diagram for detecting dynamic thresholds.
  • determining time domain feature data of the original ECG signal may include the following steps:
  • step 401 the extreme value in the wavelet coefficients is compared with the corresponding extreme threshold by a comparator.
  • Step 402 When the two consecutive extreme values are less than the corresponding extreme threshold and the time difference between the two extremes is less than the preset R wave width time, the detection of the R wave peak is triggered.
  • Step 403 extracting P waves and T waves in the original ECG signal according to the R wave peak position.
  • Step 404 determining time domain characteristic data of the original ECG signal according to the R peak, the P wave, and the T wave.
  • the time domain characteristic data of the ECG signal may include the peak position of the R wave, the peak position of the P wave, the peak position of the T wave, the amplitude value of the P wave, the amplitude value of the R wave, and the T wave.
  • the original ECG signal is subjected to wavelet transform, for example, a discrete wavelet transform of a quadratic spline. If the singular point of the original ECG signal is the intersection of a pair of rising edges and falling edges, the signal corresponding to the intersection point is After wavelet transform, it can become a zero value of a minimum value and a maximum value.
  • the R wave of the ECG signal occurs exactly at the zero-crossing position of the extreme value pair. Therefore, the R-wave peak position can be detected by the zero-crossing point of the extreme value pair of the original ECG signal on the wavelet transform.
  • the P wave and the T wave can also be extracted by the same method, and the time domain characteristic data of the ECG signal can be referred to FIG. 4B.
  • the P wave and the T wave can be detected in combination with the detection result of the previous R wave.
  • the P wave can be detected before the peak position of the R wave, for example, the P wave is detected within a period of 150 ms before the peak position of the R wave and 150 minutes before the peak position of the R wave.
  • It is also possible to detect the T wave after the peak position of the R wave for example, detecting the T wave within a period of from 400 ms after the peak position of the R wave to 400 ms after the peak position of the R wave.
  • the Q wave and the S wave can be respectively detected forward and backward based on the peak position of the R wave, and then the P wave can be detected based on the Q wave, and the S wave can be used as a reference.
  • the T wave is detected, thereby increasing the speed of detection and reducing the error rate of detection.
  • the feature data of the Q wave and the S wave in the time domain can also be used as the time domain feature data, so that the feature representation of the ECG signal in the time domain can be further improved.
  • the detection of time domain feature data of the ECG signal can be implemented by a hardware circuit.
  • the R wave peak position can be detected by a hardware circuit.
  • the first comparator 32 can compare the extreme value h of the wavelet coefficients with an extreme threshold stored in the first register 31. Taking the first comparison as an example, the value stored by default in the second register 33 can be zero. Assuming that the first comparator 32 determines that the maximum value is less than the upper threshold in the first register 31, the first comparator 32 can send a logic signal 1 to the AND gate 36. On the other hand, the first comparator 32 can also provide a logic signal to the second register 33. Based on the logic signal, the second register 33 can transmit a default value of 0 to the second comparator 35, and can also store a count value of the first counter 34, the count value being the time corresponding to the maximum value.
  • the first counter 34 may transmit the time corresponding to the maximum value to the second comparator 35.
  • the second comparator 35 may calculate a difference between the time corresponding to the maximum value and 0, and then compare the difference with a preset R wave width time.
  • the preset R wave width time can be a theoretical value, such as 0.1 s.
  • the logic signal 0 can be sent to the AND gate 36. Since the AND gate 36 receives the logic signal 1 from the first comparator 32 and the logic signal 0 from the second comparator 35, the detection of the R wave peak may not be triggered.
  • the first comparator 32 compares the minimum value with the lower threshold in the first register 31. If the minimum value is less than the lower threshold, the first comparator 32 can send a logic signal 1 to the gate 36. On the other hand, the first comparator 32 can also provide a logic signal to the second register 33. Based on the logic signal, the second register 33 may send the time corresponding to the saved maximum value to the second comparator 35, and may also store the count value of the first counter 34, the count value corresponding to the minimum value. time. The first counter 34 can transmit the time corresponding to the minimum value to the second comparator 35. The second comparator 35 may calculate a difference between the time corresponding to the maximum value and the time corresponding to the minimum value, and then compare the difference with the preset R wave width time.
  • the second comparator 35 may send a logic signal 1 to the AND gate 36 when it is determined that the difference is less than the preset R wave width time. Since the AND gate 36 receives the logic signal 1 from the first comparator 32 and the logic signal 1 from the second comparator 35, the detection of the R wave peak can be triggered.
  • the P wave and the T wave in the original ECG signal can be extracted based on the R wave peak position, thereby obtaining the time domain feature data described in the above step 404.
  • the present application can use a dynamic method to calculate the extreme threshold used in the time domain feature data, and the dynamic extremum threshold can improve the accuracy of extracting the time domain feature data.
  • the value in the third register 38 can be preset to zero; the third comparator 37 can compare the wavelet coefficients h on each scale with the values in the third register 38, respectively. If the third comparator 37 finds a larger wavelet coefficient by comparison, a logic control signal can be sent to the third register 38.
  • the third register 38 can calculate the extreme threshold as follows:
  • max1 is the maximum value of the wavelet coefficients detected during the first threshold detection period
  • min1 is the minimum value of the wavelet coefficients detected during the first threshold detection period
  • max2 is within the second threshold detection period
  • min2 is the minimum value of the wavelet coefficients detected during the second threshold detection period.
  • a+b 1
  • a and b represent the weights corresponding to the first threshold detection period and the second threshold detection period, respectively.
  • p and q are positive numbers less than one.
  • the above-calculated extreme threshold is stored in the first register 31, and the second counter 39 controls the extremum threshold in the first register 31 to be cleared when the current count reaches a threshold detection period, so as to be within the next threshold detection period. Recalculate the extreme threshold and update.
  • the detection of the time domain feature data is realized by hardware means such as a register and a comparator, and the real-time performance of detecting the time domain feature data can be improved.
  • the use of dynamic extremum thresholds can improve the accuracy of time domain feature data.
  • Combining the detection results of the R wave to detect the P wave and the T wave can improve the detection speed and reduce the detection error rate.
  • FIG. 5 illustrates a flow diagram of a wearable device-based disease detection method in accordance with yet another exemplary embodiment of the present application.
  • the disease detection method may include the following steps:
  • step 501 the user's original ECG signal is collected by the ECG sensor.
  • Step 502 determining a feature vector of the original ECG signal.
  • the feature vector may include time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal.
  • Step 503 Calculate a similarity between the feature vector and a reference feature vector of the disease ECG signal.
  • Step 504 When the maximum similarity is greater than a preset disease threshold, the disease type corresponding to the maximum similarity is identified as the disease type corresponding to the original ECG signal.
  • step 503 based on the feature vector of the original ECG signal
  • the similarity value can be calculated by the following formula:
  • a feature vector representing the user's original ECG signal Is the reference eigenvector of the ECG signal of the stored rth disease, r is a positive integer.
  • M is a matrix model obtained by a method of machine learning, and elements in the matrix model represent weight coefficients corresponding to the feature vectors.
  • T represents the transpose of the vector.
  • the wearable device is a smart bracelet
  • the smart bracelet has stored the reference feature vector corresponding to the ECG signals of disease 1 and disease 2 with Reference feature vector with It can be obtained by offline training of the ECG signals of Disease 1 and Disease 2 recorded in the case database management system provided by the relevant medical institution.
  • This embodiment can calculate the feature vector separately Reference eigenvector with The similarity values between the values are d 1 and d 2 . By finding a larger value from d 1 and d 2 and the larger value is greater than the preset disease threshold, the disease type corresponding to the larger value is used as the type of disease determined by the smart bracelet.
  • d 1 is greater than d 2 and d 1 is greater than a predetermined disease threshold
  • d 1 is greater than a predetermined disease threshold
  • the preset disease threshold can be obtained by a massive experiment.
  • the latest case report may also be generated for the user according to the currently detected disease type for the user's reference.
  • the original ECG signal detected by the electrocardiographic sensor is a normal ECG signal, and a disease-free detection report can be generated for the user's reference.
  • the identity information of the user may be authenticated by the feature vector corresponding to the original ECG signal, and then passed after the authentication is passed.
  • Steps 503 and 504 perform disease detection on the user's ECG signal.
  • the process of authenticating a user through a feature vector of the original ECG signal may be:
  • step 601 the user's original ECG signal is collected by the ECG sensor.
  • Step 602 determining a feature vector of the original ECG signal.
  • the feature vector may include time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal.
  • Step 603 Determine, according to the similarity between the feature vector and the reference feature vector of the legal user ECG signal, whether the user identity corresponding to the original ECG signal is legal.
  • step 603 a similarity between the feature vector of the current user ECG signal and the reference feature vector of the ECG signal of the legal user may be calculated; when the maximum similarity is greater than the preset identity threshold, the user may be determined to pass the identity authentication. .
  • the wearable device is a smart bracelet. If the smart bracelet has stored the reference feature vector of the user's A and user B's ECG signals. with Referring to FIG. 7, since the ECG signals of User A and User B are not the same in the time domain, the respective reference feature vectors are not the same. In this embodiment, the feature vector of the current user's ECG signal can be separately calculated. Reference eigenvector with The similarities between the two are d A and d B . If d A is greater than d B and d A is greater than a preset identity threshold, it can be confirmed that the user using the smart bracelet is User A.
  • the same user may correspond to a plurality of reference feature vectors, which may be reference feature vectors of the user during motion and at rest, respectively.
  • the feature vector of the current user's ECG signal can be calculated separately
  • the state of the user can be identified by the maximum similarity, for example, the motion state or the stationary state.
  • the present invention implements disease detection or identity authentication by means of hardware, which can effectively shorten the detection time.
  • the matrix model can be combined with the machine learning method to improve the recognition rate of ECG signals, which can ensure the accuracy of disease type detection.
  • FIG. 8 shows a schematic structural diagram of a wearable device according to an exemplary embodiment of the present application.
  • the ECG sensor 81 can acquire the original ECG signal; the signal and processing module 821 can perform wavelet transform on the original ECG signal to filter the original ECG signal to obtain wavelet coefficients; the feature vector extraction module 822 can Determining time domain feature data and frequency domain feature data; the model 823 can calculate a reference feature vector of the feature vector of the original ECG signal and the attention information (eg, a reference feature vector corresponding to the stored disease ECG signal, and/or a legitimate user) The similarity between the reference feature vectors corresponding to the ECG signals; the result discrimination module 824 can determine the state of the user's attention information according to the calculated similarity, such as determining the user's disease detection result, or determining whether the user's identity is legal.
  • the attention information eg, a reference feature vector corresponding to the stored disease ECG signal, and/or a legitimate user
  • the signal and processing module 821, the feature vector extraction module 822, the model 823, and the result discrimination module 824 may all be included in the FPGA system 82.
  • the storage module 83 can also store reference feature vectors corresponding to various types of attention information, for example, a reference feature vector corresponding to the disease ECG signal and a reference feature vector corresponding to the ECG signal of the legitimate user, thereby reducing the computational complexity of the FPGA system 82. Degree, shorten the time of ECG signal recognition, improve the efficiency of disease detection and identity authentication.
  • the present application further provides a user attention information determining apparatus based on a wearable device.
  • FIG. 9 it is a schematic diagram of a hardware structure of a device for determining user interest information based on a wearable device provided by the present application.
  • the wearable device-based user attention information determining apparatus may include a processor 910 and a machine readable storage medium 920.
  • the processor 910 and the machine readable storage medium 920 are typically interconnected by an internal bus 930.
  • the device may also include an external interface 940 to enable communication with other devices or components.
  • the machine-readable storage medium 920 may be: RAM (Radom Access Memory), volatile memory, non-volatile memory, flash memory, or the like, or a similar storage medium, or combination.
  • RAM Random Access Memory
  • volatile memory volatile memory
  • non-volatile memory non-volatile memory
  • flash memory or the like, or a similar storage medium, or combination.
  • machine-readable storage medium 920 stores machine-executable instructions corresponding to user-focused information determination logic 950 of the wearable device.
  • the determination logic 950 can include signals The acquisition unit 951, the first determination unit 952, and the second determination unit 953.
  • the signal collecting unit 951 is configured to collect the original ECG signal of the user by using the ECG sensor integrated by the wearable device;
  • a first determining unit 952 configured to determine a feature vector of the original ECG signal, where the feature vector includes time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal;
  • the second determining unit 953 is configured to determine the user's attention information based on the similarity between the feature vector and the reference feature vector indicating the attention information.
  • the attention information is health information
  • the second determining unit 953 is configured to determine a disease type corresponding to the original ECG signal according to a similarity between the feature vector and a reference feature vector of the disease ECG signal.
  • the attention information is identity information
  • the second determining unit 953 is configured to determine, according to the similarity between the feature vector and the reference feature vector of the legal user ECG signal, whether the user identity corresponding to the original ECG signal is legal.
  • the first determining unit 952 may include:
  • a wavelet transform subunit configured to perform wavelet transform on the original ECG signal to obtain a wavelet coefficient of the original ECG signal
  • a first determining subunit configured to determine the wavelet coefficient as a frequency domain characteristic data of the original ECG signal.
  • the first determining unit 952 may further include:
  • a first operation subunit configured to perform autocorrelation operation on the wavelet transform filtered electrocardiographic signal, and discrete cosine transform to obtain discrete cosine transform coefficients of the autocorrelation and discrete cosine transform;
  • a second determining subunit configured to determine the discrete cosine transform coefficient as a frequency domain characteristic data of the original ECG signal.
  • the first determining unit 952 may further include:
  • Detecting a triggering subunit configured to trigger detection of an R wave peak when two consecutive extreme values are less than a corresponding extreme threshold and the time difference of the two extremes is less than a preset R wave width time;
  • a third determining subunit configured to extract P waves and T waves in the original ECG signal according to the R wave peak position
  • a fourth determining subunit configured to determine time domain characteristic data of the original ECG signal according to the R peak, the P wave, and the T wave, where the time domain characteristic data includes: a peak of the R wave a position, a peak position of the P wave, a peak position of the T wave, an amplitude value of the P wave, an amplitude value of the R wave, an amplitude value of the T wave, and a peak position of the P wave.
  • the second determining unit 953 may include:
  • a calculation subunit configured to calculate a similarity between the feature vector and a reference feature vector representing the attention information
  • a fifth determining subunit configured to determine the attention information corresponding to the maximum similarity as the attention information corresponding to the original ECG signal when the maximum similarity is greater than the preset attention threshold.
  • the similarity value can be calculated by the following formula:
  • a feature vector representing the user's original ECG signal For the stored rth reference feature vector, r is a positive integer, and M is a matrix model obtained by a machine learning method, and elements in the matrix model represent weight coefficients corresponding to the feature vector.
  • the determination of how the wearable device performs the determination logic 950 is described below by way of a software implementation.
  • the user interest information determination logic 950 of the present application based on the wearable device should be understood to be computer instructions stored in the machine readable storage medium 920.
  • the processor 910 on the wearable device of the present application executes the logic 950, the processor 910 performs the following operations by calling an instruction corresponding to the determination logic 950 stored on the machine readable storage medium 920:
  • Determining a feature vector of the original ECG signal the feature vector including time domain feature data of the original ECG signal and frequency domain feature data of the original ECG signal;
  • the user's attention information is determined based on the similarity between the feature vector and the reference feature vector representing the attention information.
  • the attention information is health information
  • the information executable by the machine causes the processor to:
  • the attention information is identity information
  • the machine executable instructions cause the processor to: upon determining a user's attention information based on a similarity between the feature vector and a reference feature vector representing the attention information:
  • the machine executable instructions cause the processor to: upon determining a feature vector of the original ECG signal:
  • the wavelet coefficients are determined as a frequency domain characteristic data of the original electrocardiographic signal.
  • the machine executable instructions cause the processor to: upon determining a feature vector of the original ECG signal:
  • the discrete cosine transform coefficients are determined as a frequency domain characteristic data of the original electrocardiographic signal.
  • the machine executable instructions cause the processor to: upon determining a feature vector of the original ECG signal:
  • time domain characteristic data of the original ECG signal according to the R peak, the P wave, and the T wave
  • the time domain characteristic data comprising: a peak position of the R wave, a peak of the P wave a position, a peak position of the T wave, an amplitude value of the P wave, an amplitude value of the R wave, an amplitude value of the T wave, an interval between a peak position of the P and a peak position of the R wave.
  • the machine executable instructions cause the processor to: upon determining a user's attention information based on a similarity between the feature vector and a reference feature vector representing the attention information:
  • the attention information corresponding to the maximum similarity is determined as the attention information corresponding to the original ECG signal.
  • the similarity value can be calculated by the following formula:
  • a feature vector representing the user's original ECG signal For the stored rth reference feature vector, r is a positive integer, and M is a matrix model obtained by a machine learning method, and elements in the matrix model represent weight coefficients corresponding to the feature vector.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Cardiology (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Theoretical Computer Science (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Physiology (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Signal Processing (AREA)
  • Algebra (AREA)
  • Computer Security & Cryptography (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychiatry (AREA)
  • General Business, Economics & Management (AREA)
  • Business, Economics & Management (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)

Abstract

一种基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备，该方法包括：通过可穿戴设备集成的心电传感器（81）采集用户的原始心电信号（101）；确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据（102）；基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息（103）。

Description

基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备 技术领域
本申请涉及可穿戴设备技术领域。
背景技术
随着社会的发展和生活水平的提高，以及移动互联网的快速发展，可穿戴设备以其穿戴便捷的优势走入了人们的生活。
发明内容
根据本申请的第一方面，提出了一种基于可穿戴设备的用户关注信息确定方法，应用在可穿戴设备上，包括：
通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号对应的时域特征数据和所述原始心电信号的频域特征数据；
基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
根据本申请的第二方面，提出了一种可穿戴设备，包括：
心电传感器，用于采集用户的原始心电信号；以及
FPGA系统，用于基于所述原始心电信号确定用户的关注信息，包括：
特征向量提取模块，用于确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
结果判别模块，用于基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
根据本申请的第三方面，提出了一种基于可穿戴设备的用户关注信息确定装置，应用在可穿戴设备上，包括：
处理器；
用于存储机器可执行指令的存储器；
其中，通过读取并执行所述存储器所存储的与基于可穿戴设备的用户关注信息确定逻辑对应的机器可执行指令，所述处理器被促使：
通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
由以上技术方案可见，本申请可通过心电传感器采集用户的原始心电信号，确定原始心电信号的特征向量，可基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。由于所述特征向量可包括原始心电信号的时域特征数据和频域特征数据，相似度算法中所使用的矩阵模型可以通过机器学习的方法得到，可使得通过ECG确定的用户关注信息的准确度较高。
附图说明
图1示出了根据本申请的一示例性实施例的基于可穿戴设备的用户关注信息确定方法的流程示意图。
图2A示出了根据本申请的一示例性实施例的基于可穿戴设备的疾病检测方法的流程示意图；
图2B示出了根据本申请的一示例性实施例的原始心电信号的示意图；
图3A示出了根据本申请的一示例性实施例的确定原始心电信号的频域特征数据的流程示意图；
图3B示出了根据本申请的一示例性实施例的通过小波变换滤除噪声后的心电信号的示意图；
图4A示出了根据本申请的一示例性实施例的确定原始心电信号的时域特征数据的流程示意图；
图4B示出了心电信号的时序和幅值特征的示意图；
图4C示出了采用硬件电路检测R波波峰的示意图；
图4D示出了用于检测动态阈值的电路图；
图5示出了根据本发明的又一示例性实施例的基于可穿戴设备的疾病检测方法的流程示意图；
图6示出了根据本申请的一示例性实施例的基于可穿戴设备的身份识别方法的流程示意图；
图7示出了两个不同用户的心电信号的示意图；
图8示出了根据本申请的一示例性实施例的可穿戴设备的结构示意图；
图9示出了根据本申请的一示例性实施例的基于可穿戴设备的用户关注信息确定装置的硬件结构示意图；
图10示出了根据本申请的一示例性实施例的基于可穿戴设备的用户关注信息确定逻辑的功能模块框图；
图11示出了根据本申请的又一示例性实施例的基于可穿戴设备的用户关注信息确定逻辑的功能模块框图。
具体实施方式
图1示出了根据本申请的一示例性实施例的基于可穿戴设备的用户关注信息确定方法的流程示意图。参见图1，该方法可以包括：
步骤101，通过可穿戴设备集成的心电传感器采集用户的原始心电信号。
步骤102，确定所述原始心电信号的特征向量，所述特征向量可包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据。
步骤103，基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
在本例中，所述关注信息可以为健康信息，包括用户是否患有某种疾病。所述关注信息也可以为身份信息，包括是否为合法用户等。
下面结合具体的例子来描述本申请的实现过程。
与指纹、人脸图像、虹膜等生物特征相比，人体的心电信号(Electrocardiograph，ECG)可由个体的心脏结构决定，其具有普适性、唯一性、易采集、永久性等特点。并且，ECG还具有只存在于活体，不易被仿制，不易丢失等优势。因此，可通过可穿戴设备采集ECG，并基于ECG对用户的心血管疾病进行检测，能够使用户及时发现身体出现的异常，确保用户能 够及时就医治疗。
本申请可通过可穿戴设备中集成的心电传感器采集用户的原始心电信号，可确定原始心电信号的特征向量，根据所述特征向量与疾病心电信号的参考特征向量的相似度确定原始心电信号对应的疾病类型。由于特征向量包括原始心电信号的时域特征数据和频域特征数据，通过基于ECG的特征向量对用户进行疾病检测，可以大大提高用户疾病检测的准确度，从而可基于可穿戴设备实现对疾病的尽早识别。
为对本申请进行说明，提供下列实施例：
图2A示出了根据本申请的一示例性实施例的基于可穿戴设备的疾病检测方法的流程示意图。图2B示出了根据本申请的一示例性实施例的原始心电信号的示意图。所述基于可穿戴设备的疾病检测方法可以应用到可穿戴设备上，例如，智能手环等设备上，所述可穿戴设备上可以设置有心电传感器。参考图2A，所述疾病检测方法可包括如下步骤：
步骤201，通过心电传感器采集用户的原始心电信号。
步骤202，确定所述原始心电信号的特征向量。其中，所述特征向量可包括原始心电信号的时域特征数据和原始心电信号的频域特征数据。
步骤203，基于所述特征向量与疾病心电信号的参考特征向量的相似度确定原始心电信号对应的疾病类型。
在步骤201中，参考图2B，心电传感器采集到的原始心电信号通常有较强的噪声，噪声可随着时间的推移而变化，但是同一个用户不同时刻采集的原始心电信号的QRS波群、P波、T波基本相同。
在步骤202中，在一实施例中，原始心电信号的频域特征数据可以包括原始心电信号对应的小波系数、离散余弦变换系数、傅立叶变换系数、HHT(Hilbert-Hwang)变换系数等，本申请对具体的变换方式不做限制。
在步骤203中，在一实施例中，可以对已有用户的心电数据进行训练，以获得各种疾病的心电信号的参考特征向量。在一实施例中，疾病心电信号的参考特征向量可以在线下训练完成，在实现疾病检测时可直接使用训练得到的疾病心电信号的参考特征向量即可。例如最近邻算法的相似度算法中所使用的矩阵模型可以通过机器学习的方法得到，机器学习的目标以相似度算法的准确率尽量高为准，从而可以确保疾病检测的准确度。
由上述描述可知，本申请实施例通过心电传感器采集用户的原始心电信号，确定原始心 电信号的特征向量，并可根据例如最近邻算法等的相似度算法来计算所述特征向量与各种疾病心电信号的参考特征向量的相似度，从而可根据所述相似度来确定原始心电信号对应的疾病类型。由于特征向量可包括原始心电信号的时域特征数据和频域特征数据，相似度算法中所使用的矩阵模型可以通过机器学习的方法得到，以使得通过ECG对用户进行疾病检测的准确度较高。
图3A示出了根据本申请的一示例性实施例的确定原始心电信号的频域特征数据的流程示意图。图3B示出了根据本申请的一示例性实施例的通过小波变换滤除噪声后的心电信号的示意图。参考图3A，确定原始心电信号的频域特征数据可以包括如下步骤：
步骤301，对原始心电信号进行小波变换，得到原始心电信号的小波系数。
步骤302，将所述小波系数确定为原始心电信号的一种频域特征数据。
步骤303，对小波变换后的心电信号进行自相关，和离散余弦变换，得到自相关和离散余弦变换后的离散余弦变换系数。
步骤304，将所述离散余弦变换系数确定为原始心电信号的另一种频域特征数据。
在步骤301中，通过小波变换可以将原始心电信号中不同频率的信号分解出来。由于在低尺度上主要反映原始心电信号的高频噪声，在高尺度上主要反映原始心电信号的低频噪声，采用小波变换后的中间尺度对原始心电信号进行分析，可以有效地将有用信号和干扰信号区分开。在一实施例中，可以通过一组系数预定的高通和低通滤波器对原始心电信号进行小波分解，从而得到原始心电信号对应的小波系数。小波系数可以包括多级尺度上的小波系数。在一实施例中，可以通过FPGA的硬件平台进行移位和加法的方式实现对原始心电信号的小波变换，而移位和加法运算的逻辑简单，易于实现。参见图3B，小波变换后，原始心电信号的噪声可得到有效去除，得到的心电信号更加有规律。
在步骤303中，在一实施例中，可以对滤波后的心电信号进行自相关运算，通过自相关运算可以消除心电信号中与识别无关的信号部分。接着，可再将自相关运算之后的心电信号进行离散余弦变换，进而得到离散余弦变换系数。在另一实施例中，还可以对滤波后的心电信号进行傅立叶变换或者HHT变换等，将变换后的傅立叶变换系数或HHT变换系数作为心电信号的一种频域特征数据。
本实施例通过小波变换可滤除原始心电信号的噪声，可使得心电信号更加有规律，从而可确保原始心电信号在各尺度上的小波系数和离散余弦变换系数更加准确地表示原始心电信号在频域的特征。
图4A示出了根据本申请的一示例性实施例的确定原始心电信号的时域特征数据的流程示意图。图4B示出了心电信号的时序和幅值特征的示意图。图4C示出了采用硬件电路检测R波波峰的示意图。图4D示出了用于检测动态阈值的电路图。参考图4A，确定原始心电信号的时域特征数据可以包括如下步骤：
步骤401，通过比较器将小波系数中的极值与对应的极值阈值进行比较。
步骤402，当连续的两个极值小于对应的极值阈值并且所述两个极值的时间差值小于预设的R波宽度时间时，触发R波波峰的检测。
步骤403，根据R波波峰位置提取原始心电信号中的P波和T波。
步骤404，根据R波峰、P波、T波确定原始心电信号的时域特征数据。
其中，如图4B所示，心电信号的时域特征数据可包括R波的波峰位置、P波的波峰位置、T波的波峰位置、P波的幅度值、R波的幅度值、T波的幅度值，P波与R波峰的间隔、T波与R波峰的间隔、PR段、ST段。
由参见图4B，对原始心电信号进行小波变换，例如二次样条的离散小波变换，原始心电信号的奇异点如果是一对上升沿和下降沿的交点，则该交点对应的信号经小波变换后可成为一个极小值和一个极大值的过零点。而心电信号的R波发生的位置正好是极值对的过零点位置，因此可通过原始心电信号在小波变换上的极值对的过零点检测出R波波峰位置。P波和T波也可以通过相同的方法提取，心电信号在的时域特征数据可以参考图4B。
在一实施例中，可以结合前一个R波的检测结果来检测P波和T波。例如，在确定R波的波峰位置后，可在R波的波峰位置之前检测P波，例如在R波的波峰位置前250ms至R波的波峰位置前150ms的时间段内检测P波。还可在R波的波峰位置之后检测T波，例如在R波的波峰位置后170ms至R波的波峰位置后400ms的时间段内检测T波。
在另一实施例中，由于PQRST波是连续的，可以R波的波峰位置为基准向前向后分别检测到Q波、S波，接着可以Q波为基准检测P波，可以S波为基准检测T波，从而可以提高检测的速度，并降低检测的错误率。
在另一实施例中，还可以将Q波和S波在时域上的特征数据作为所述时域特征数据，从而可以进一步提高心电信号在时域中的特征表示。
在一实施例中，可以通过硬件电路实现心电信号的时域特征数据的检测。参考图4C，可以通过硬件电路检测R波波峰位置。
在一个例子中，第一比较器32可将小波系数中的极值h与存储在第一寄存器31中的极值阈值进行比较。以首次比较为例，第二寄存器33中缺省存储的值可为0。假设第一比较器32确定极大值小于第一寄存器31中的上限阈值，第一比较器32可发送逻辑信号1给与门36。另一方面，第一比较器32还可为第二寄存器33提供逻辑信号。基于所述逻辑信号，第二寄存器33可将缺省值0发送给第二比较器35，还可存储第一计数器34的计数值，该计数值为所述极大值对应的时间。第一计数器34可将所述极大值对应的时间发送给第二比较器35。第二比较器35可计算所述极大值对应的时间与0的差值，然后将该差值与预设的R波宽度时间进行比较。该预设的R波宽度时间可为理论值，例如0.1s。第二比较器35确定该差值大于预设的R波宽度时间时，可向与门36发送逻辑信号0。由于与门36接收到来自第一比较器32的逻辑信号1以及来自第二比较器35的逻辑信号0，可不触发R波波峰的检测。
接着，第一比较器32将极小值与第一寄存器31中的下限阈值进行比较。若所述极小值小于所述下限阈值，第一比较器32可发送逻辑信号1给与门36。另一方面，第一比较器32还可为第二寄存器33提供逻辑信号。基于所述逻辑信号，第二寄存器33可将保存的极大值对应的时间发送给第二比较器35，还可存储第一计数器34的计数值，该计数值为所述极小值对应的时间。第一计数器34可将所述极小值对应的时间发送给第二比较器35。第二比较器35可计算所述极大值对应的时间与所述极小值对应的时间的差值，然后将该差值与所述预设的R波宽度时间进行比较。第二比较器35在确定该差值小于所述预设的R波宽度时间时，可向与门36发送逻辑信号1。由于与门36接收到来自第一比较器32的逻辑信号1以及来自第二比较器35的逻辑信号1，可触发R波波峰的检测。
在一实施例中，可以基于R波波峰位置提取原始心电信号中的P波和T波，进而得到上述步骤404中所述的时域特征数据。
由于心电信号的波动、基漂等因素可导致QRS波在不同时刻的幅度值不同，可使得同一时刻不同尺度上和同一尺度不同时刻的极大值和极小值都不相同。因此，本申请可以采用动态的方法计算时域特征数据所使用的极值阈值，通过动态极值阈值可以提高提取时域特征数据的准确性。参考图4D，第三寄存器38中的值可以预置为0；第三比较器37可分别将各尺度上的小波系数h与第三寄存器38中的值进行比较。若第三比较器37经比较发现较大的小波系数，则可向第三寄存器38发送一个逻辑控制信号。第三寄存器38可按照如下方法计算得到的极值阈值：
上限阈值：MAX＝＝p(a*max1+b*max2)；
下限阈值：MIN＝＝q(a*min1+b*min2)。
其中，max1为在第一个阈值检测周期内检测到的小波系数的最大值，min1为在第一个阈值检测周期内检测到的小波系数的最小值，max2为在第二个阈值检测周期内检测到的小波系数的最大值，min2为在第二个阈值检测周期内检测到的小波系数的最小值。a+b＝1，a和b分别表示第一个阈值检测周期和第二阈值检测周期对应的权重。p和q为小于1的正数。
将上述计算得到的极值阈值存储在第一寄存器31中，第二计数器39控制当前计数达到一个阈值检测周期时，对第一寄存器31中的极值阈值清零，以便在下一个阈值检测周期内重新计算极值阈值并更新。
本实施例中，通过寄存器、比较器等硬件方式实现了时域特征数据的检测，可提高检测时域特征数据的实时性。采用动态极值阈值可以提高时域特征数据的精度。结合R波的检测结果来检测P波和T波可以提高检测的速度，并降低检测的错误率。
图5示出了根据本申请的又一示例性实施例的基于可穿戴设备的疾病检测方法的流程示意图。如图5所示，该疾病检测方法可包括如下步骤：
步骤501，通过心电传感器采集用户的原始心电信号。
步骤502，确定所述原始心电信号的特征向量。其中，特征向量可包括原始心电信号的时域特征数据和原始心电信号的频域特征数据。
步骤503，计算所述特征向量与疾病心电信号的参考特征向量之间的相似度。
步骤504，当最大相似度大于预设的疾病阈值时，将最大相似度对应的疾病类型识别为原始心电信号对应的疾病类型。
在步骤503中，基于原始心电信号的特征向量
Figure PCTCN2016105720-appb-000001
可以通过如下公式计算得到相似度值：
Figure PCTCN2016105720-appb-000002
其中，
Figure PCTCN2016105720-appb-000003
表示用户的原始心电信号的特征向量。
Figure PCTCN2016105720-appb-000004
为已存储的第r个疾病的心电信号的参考特征向量，r为正整数。M为通过机器学习的方法得到的矩阵模型，所述矩阵模型中的元素表示所述特征向量对应的权重系数。T表示向量的转置。
假设，可穿戴设备为智能手环，如果智能手环中已经存储了疾病1和疾病2的心电信号对应的参考特征向量
Figure PCTCN2016105720-appb-000005
和
Figure PCTCN2016105720-appb-000006
参考特征向量
Figure PCTCN2016105720-appb-000007
和
Figure PCTCN2016105720-appb-000008
可以通过对相关医疗机构提供的病例数据库管理系统中记录的疾病1和疾病2的心电信号进行线下训练得到。本实施例可以分别计算 特征向量
Figure PCTCN2016105720-appb-000009
与参考特征向量
Figure PCTCN2016105720-appb-000010
和
Figure PCTCN2016105720-appb-000011
间的相似度值为d1和d2。通过从d1和d2找到较大值，并且该较大值大于预设的疾病阈值的情形下，将该较大值对应的疾病类型作为智能手环确定出的疾病类型。例如，d1大于d2，并且d1大于预设的疾病阈值，则可以确定用户可能患有参考特征向量
Figure PCTCN2016105720-appb-000012
对应的疾病。其中，所述预设疾病阈值可以通过海量的试验来得到。
在一实施例中，还可以根据当前检测到的疾病类型为用户生成最新的病例报告，以供用户参考。
在一实施例中，如果最大相似度值小于预设的疾病阈值，可说明心电传感器检测到的原始心电信号为正常的心电信号，可以生成一个无疾病的检测报告供用户参考。
在一实施例中，为了防止智能手环被非法用户窃取所导致的合法用户的隐私信息泄露，可以通过原始心电信号对应的特征向量对用户的身份信息进行认证，在认证通过后，再通过步骤503和步骤504对用户的心电信号进行疾病检测。
参见图6，通过原始心电信号的特征向量对用户进行身份认证的过程可以为：
步骤601，通过心电传感器采集用户的原始心电信号。
步骤602，确定所述原始心电信号的特征向量。其中，所述特征向量可包括原始心电信号的时域特征数据和原始心电信号的频域特征数据。
步骤603，根据所述特征向量与合法用户心电信号的参考特征向量的相似度确定原始心电信号对应的用户身份是否合法。
在步骤603中，可计算当前用户心电信号的特征向量与合法用户的心电信号的参考特征向量之间的相似度；当最大相似度大于预设的身份阈值时，可确定用户通过身份认证。
假设，可穿戴设备为智能手环，如果智能手环中已经存储了用户A和用户B的心电信号的参考特征向量
Figure PCTCN2016105720-appb-000013
和
Figure PCTCN2016105720-appb-000014
请参考图7，由于用户A和用户B的心电信号在时域的形状并不相同，因此各自的参考特征向量也不会相同。本实施例可以分别计算当前用户的心电信号的特征向量
Figure PCTCN2016105720-appb-000015
与参考特征向量
Figure PCTCN2016105720-appb-000016
和
Figure PCTCN2016105720-appb-000017
间的相似度为dA和dB。如果dA大于dB，且dA大于预设的身份阈值，则可以确认使用智能手环的用户为用户A。
在一实施例中，同一个用户可以对应多个参考特征向量，多个参考特征向量可分别为用户在运动以及静止时的参考特征向量。例如，可分别计算当前用户的心电信号的特征向量
Figure PCTCN2016105720-appb-000018
与 同一个用户的多个特征参考向量的相似度，通过最大相似度可识别出用户的状态，例如，运动状态或者静止状态等。
本领域技术人员可以理解的是，进行疾病检测和身份认证的相似度算法中所采用的相关参数可以不同，也可以相同。
本申请通过硬件的方式实现疾病检测或身份认证，可有效缩短检测用时。此外，结合机器学习的方法得到矩阵模型可以提高心电信号的识别率，进而可以确保疾病类型检测的准确度。
图8示出了根据本申请的一示例性实施例的可穿戴设备的结构示意图。参考图8，心电传感器81可采集得到原始心电信号；信号与处理模块821可对原始心电信号进行小波变换进而对原始心电信号进行滤波处理，得到小波系数；特征向量提取模块822可确定时域特征数据和频域特征数据；模型823可计算原始心电信号的特征向量与关注信息的参考特征向量(例如，已存储的疾病心电信号对应的参考特征向量，和/或合法用户心电信号对应的参考特征向量)之间的相似度；结果判别模块824可根据所计算出的相似度确定用户的关注信息状态，例如确定用户的疾病检测结果，或者确定用户的身份是否合法。其中，信号与处理模块821、特征向量提取模块822、模型823、结果判别模块824均可包括在FPGA系统82中。存储模块83还可存储有各种关注信息对应的参考特征向量，例如包括疾病心电信号对应的参考特征向量以及合法用户的心电信号对应的参考特征向量，从而降低了FPGA系统82的计算复杂度，缩短心电信号识别的时间，提高了疾病检测以及身份认证的效率。
对应于上述方法，本申请还提供了一种基于可穿戴设备的用户关注信息确定装置。参见图9，该图示出了本申请提供的基于可穿戴设备的用户关注信息确定装置的硬件结构示意图。如图9所示，所述基于可穿戴设备的用户关注信息确定装置可包括处理器910以及机器可读存储介质920。其中，处理器910和机器可读存储介质920通常借由内部总线930相互连接。在其他可能的实现方式中，所述装置还可能包括外部接口940，以能够与其他设备或者部件进行通信。
在不同的例子中，所述机器可读存储介质920可以是：RAM(Radom Access Memory，随机存取存储器)、易失存储器、非易失性存储器、闪存，或者类似的存储介质，或者它们的组合。
进一步地，如图10所示，机器可读存储介质920上存储有基于可穿戴设备的用户关注信息确定逻辑950对应的机器可执行指令。从功能上划分，所述确定逻辑950可包括信号 采集单元951、第一确定单元952和第二确定单元953。
其中，信号采集单元951，用于通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
第一确定单元952，用于确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
第二确定单元953，用于基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
根据一个例子，所述关注信息为健康信息，
所述第二确定单元953，可用于根据所述特征向量与疾病心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的疾病类型。
根据另一个例子，所述关注信息为身份信息，
所述第二确定单元953，可用于根据所述特征向量与合法用户心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的用户身份是否合法。
根据另一个例子，所述第一确定单元952，可包括：
小波变换子单元，用于对所述原始心电信号进行小波变换，得到所述原始心电信号的小波系数；
第一确定子单元，用于将所述小波系数确定为所述原始心电信号的一种频域特征数据。
根据另一个例子，所述第一确定单元952，还可包括：
第一运算子单元，用于对所述小波变换滤波后的心电信号进行自相关运算，和离散余弦变换，得到所述自相关和离散余弦变换后的离散余弦变换系数；
第二确定子单元，用于将所述离散余弦变换系数确定为所述原始心电信号的一种频域特征数据。
根据另一个例子，所述第一确定单元952，还可包括：
比较子单元，用于通过比较器将小波系数中的极值与对应的极值阈值进行比较；
检测触发子单元，用于当连续的两个极值均小于对应的极值阈值并且所述两个极值的时间差值小于预设的R波宽度时间时，触发R波波峰的检测；
第三确定子单元，用于根据所述R波波峰位置提取所述原始心电信号中的P波和T波；
第四确定子单元，用于根据所述R波峰、所述P波、所述T波确定所述原始心电信号的时域特征数据，所述时域特征数据包括：所述R波的波峰位置、所述P波的波峰位置、所述T波的波峰位置、所述P波的幅度值、所述R波的幅度值、所述T波的幅度值，所述P的波峰位置与所述R波的波峰位置的间隔、所述T波的波峰位置与所述R波的波峰位置的间隔、PR段、ST段。
根据另一个例子，所述第二确定单元953，可包括：
计算子单元，用于计算所述特征向量与表示关注信息的参考特征向量之间的相似度；
第五确定子单元，用于当最大相似度大于预设的关注阈值时，将最大相似度对应的关注信息确定为原始心电信号对应的关注信息。
根据另一个例子，可通过如下公式计算所述相似度值：
Figure PCTCN2016105720-appb-000019
其中，
Figure PCTCN2016105720-appb-000020
表示用户的原始心电信号的特征向量，
Figure PCTCN2016105720-appb-000021
为已存储的第r个参考特征向量，r为正整数，M为通过机器学习的方法得到的矩阵模型，所述矩阵模型中的元素表示所述特征向量对应的权重系数。
下面以软件实现为例，描述可穿戴设备如何执行该确定逻辑950。在该例子中，本申请基于可穿戴设备的用户关注信息确定逻辑950应理解为存储在机器可读存储介质920中的计算机指令。当本申请可穿戴设备上的处理器910执行该逻辑950时，该处理器910通过调用机器可读存储介质920上保存的确定逻辑950对应的指令执行如下操作：
通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
根据一个例子，所述关注信息为健康信息，
在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注 信息时，所述机器可执行指令促使所述处理器：
根据所述特征向量与疾病心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的疾病类型。
根据另一个例子，所述关注信息为身份信息，
在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息时，所述机器可执行指令促使所述处理器：
根据所述特征向量与合法用户心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的用户身份是否合法。
根据另一个例子，在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
对所述原始心电信号进行小波变换，得到所述原始心电信号的小波系数；
将所述小波系数确定为所述原始心电信号的一种频域特征数据。
根据另一个例子，在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
对所述小波变换滤波后的心电信号进行自相关运算，和离散余弦变换，得到所述自相关和离散余弦变换后的离散余弦变换系数；
将所述离散余弦变换系数确定为所述原始心电信号的一种频域特征数据。
根据另一个例子，在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
通过比较器将小波系数中的极值与对应的极值阈值进行比较；
当连续的两个极值均小于对应的极值阈值并且所述两个极值的时间差值小于预设的R波宽度时间时，触发R波波峰的检测；
根据所述R波波峰位置提取所述原始心电信号中的P波和T波；
根据所述R波峰、所述P波、所述T波确定所述原始心电信号的时域特征数据，所述时域特征数据包括：所述R波的波峰位置、所述P波的波峰位置、所述T波的波峰位置、所述P波的幅度值、所述R波的幅度值、所述T波的幅度值，所述P的波峰位置与所述R波的波峰位置的间隔、所述T波的波峰位置与所述R波的波峰位置的间隔、PR段、ST段。
根据另一个例子，在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息时，所述机器可执行指令促使所述处理器：
计算所述特征向量与表示关注信息的参考特征向量之间的相似度；
当最大相似度大于预设的关注阈值时，将最大相似度对应的关注信息确定为原始心电信号对应的关注信息。
根据另一个例子，可通过如下公式计算所述相似度值：
Figure PCTCN2016105720-appb-000022
其中，
Figure PCTCN2016105720-appb-000023
表示用户的原始心电信号的特征向量，
Figure PCTCN2016105720-appb-000024
为已存储的第r个参考特征向量，r为正整数，M为通过机器学习的方法得到的矩阵模型，所述矩阵模型中的元素表示所述特征向量对应的权重系数。

Claims (18)

  1. 一种基于可穿戴设备的用户关注信息确定方法，包括：
    通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
    确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
    基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
  2. 根据权利要求1所述的方法，其中，
    所述关注信息为健康信息，
    所述基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息，包括：根据所述特征向量与疾病心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的疾病类型。
  3. 根据权利要求1所述的方法，其中，
    所述关注信息为身份信息，
    所述基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息，包括：根据所述特征向量与合法用户心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的用户身份是否合法。
  4. 根据权利要求1所述的方法，其中，所述确定所述原始心电信号的特征向量，包括：
    对所述原始心电信号进行小波变换，得到所述原始心电信号的小波系数；
    将所述小波系数确定为所述原始心电信号的一种频域特征数据。
  5. 根据权利要求4所述的方法，其中，所述确定所述原始心电信号的特征向量，包括：
    对所述小波变换滤波后的心电信号进行自相关运算，和离散余弦变换，得到所述自相关和离散余弦变换后的离散余弦变换系数；
    将所述离散余弦变换系数确定为所述原始心电信号的一种频域特征数据。
  6. 根据权利要求4所述的方法，其中，所述确定所述原始心电信号的特征向量，包括：
    通过比较器将小波系数中的极值与对应的极值阈值进行比较；
    当连续的两个极值均小于对应的极值阈值并且所述两个极值的时间差值小于预设的R波宽度时间时，触发R波波峰的检测；
    根据所述R波波峰位置提取所述原始心电信号中的P波和T波；
    根据所述R波峰、所述P波、所述T波确定所述原始心电信号的时域特征数据，所述时域特征数据包括：所述R波的波峰位置、所述P波的波峰位置、所述T波的波峰位置、所述P波的幅度值、所述R波的幅度值、所述T波的幅度值，所述P的波峰位置与所述R波的波峰位置的间隔、所述T波的波峰位置与所述R波的波峰位置的间隔、PR段、ST段。
  7. 根据权利要求1所述的方法，其中，基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息，包括：
    计算所述特征向量与表示关注信息的参考特征向量之间的相似度；
    当最大相似度大于预设的关注阈值时，将最大相似度对应的关注信息确定为原始心电信号对应的关注信息。
  8. 根据权利要求7所述的方法，其中，通过如下公式计算所述相似度值：
    Figure PCTCN2016105720-appb-100001
    其中，
    Figure PCTCN2016105720-appb-100002
    表示用户的原始心电信号的特征向量，
    Figure PCTCN2016105720-appb-100003
    为已存储的第r个参考特征向量，r为正整数，M为通过机器学习的方法得到的矩阵模型，所述矩阵模型中的元素表示所述特征向量对应的权重系数。
  9. 一种可穿戴设备，包括：
    心电传感器，用于采集用户的原始心电信号；以及
    FPGA系统，用于基于所述原始心电信号确定用户的关注信息，包括：
    特征向量提取模块，用于确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
    结果判别模块，用于基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
  10. 根据权利要求9所述的可穿戴设备，其中，
    所述FPGA系统还包括：
    信号与处理模块，用于对原始心电信号进行小波变换进而对原始心电信号进行滤波处理，得到小波系数；
    模型，用于计算原始心电信号的特征向量与关注信息的参考特征向量之间的相似度。
  11. 一种基于可穿戴设备的用户关注信息确定装置，包括：
    处理器；
    用于存储机器可执行指令的存储器；
    其中，通过读取并执行所述存储器所存储的与基于可穿戴设备的用户关注信息确定逻辑对应的机器可执行指令，所述处理器被促使：
    通过所述可穿戴设备集成的心电传感器采集用户的原始心电信号；
    确定所述原始心电信号的特征向量，所述特征向量包括所述原始心电信号的时域特征数据和所述原始心电信号的频域特征数据；
    基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息。
  12. 根据权利要求11所述的装置，其中，
    所述关注信息为健康信息，
    在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息时，所述机器可执行指令促使所述处理器：
    根据所述特征向量与疾病心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的疾病类型。
  13. 根据权利要求11所述的装置，其中，
    所述关注信息为身份信息，
    在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息时，所述机器可执行指令促使所述处理器：
    根据所述特征向量与合法用户心电信号的参考特征向量之间的相似度，确定所述原始心电信号对应的用户身份是否合法。
  14. 根据权利要求11所述的装置，其中，
    在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
    对所述原始心电信号进行小波变换，得到所述原始心电信号的小波系数；
    将所述小波系数确定为所述原始心电信号的一种频域特征数据。
  15. 根据权利要求14所述的装置，其中，
    在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
    对所述小波变换滤波后的心电信号进行自相关运算，和离散余弦变换，得到所述自相关和离散余弦变换后的离散余弦变换系数；
    将所述离散余弦变换系数确定为所述原始心电信号的一种频域特征数据。
  16. 根据权利要求14所述的装置，其中，
    在确定所述原始心电信号的特征向量时，所述机器可执行指令促使所述处理器：
    通过比较器将小波系数中的极值与对应的极值阈值进行比较；
    当连续的两个极值均小于对应的极值阈值并且所述两个极值的时间差值小于预设的R波宽度时间时，触发R波波峰的检测；
    根据所述R波波峰位置提取所述原始心电信号中的P波和T波；
    根据所述R波峰、所述P波、所述T波确定所述原始心电信号的时域特征数据，所述时域特征数据包括：所述R波的波峰位置、所述P波的波峰位置、所述T波的波峰位置、所述P波的幅度值、所述R波的幅度值、所述T波的幅度值，所述P的波峰位置与所述R波的波峰位置的间隔、所述T波的波峰位置与所述R波的波峰位置的间隔、PR段、ST段。
  17. 根据权利要求11所述的装置，其中，
    在基于所述特征向量和表示关注信息的参考特征向量之间的相似度，确定用户的关注信息时，所述机器可执行指令促使所述处理器：
    计算所述特征向量与表示关注信息的参考特征向量之间的相似度；
    当最大相似度大于预设的关注阈值时，将最大相似度对应的关注信息确定为原始心电信号对应的关注信息。
  18. 根据权利要求17所述的装置，其中，通过如下公式计算所述相似度值：
    Figure PCTCN2016105720-appb-100004
    其中，
    Figure PCTCN2016105720-appb-100005
    表示用户的原始心电信号的特征向量，
    Figure PCTCN2016105720-appb-100006
    为已存储的第r个参考特征向量，r为正整数，M为通过机器学习的方法得到的矩阵模型，所述矩阵模型中的元素表示所述特征向量对应的权重系数。
PCT/CN2016/105720 2015-11-17 2016-11-14 基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备 WO2017084546A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/584,911 US10163528B2 (en) 2015-11-17 2017-05-02 Determining user-interested information based on wearable device

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201510795936.2A CN105286853B (zh) 2015-11-17 2015-11-17 基于可穿戴设备的疾病检测方法及装置、可穿戴设备
CN201510796544.8A CN105468951B (zh) 2015-11-17 2015-11-17 通过心电特征进行身份识别的方法及装置、可穿戴设备
CN201510796544.8 2015-11-17
CN201510795936.2 2015-11-17

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/584,911 Continuation-In-Part US10163528B2 (en) 2015-11-17 2017-05-02 Determining user-interested information based on wearable device

Publications (1)

Publication Number Publication Date
WO2017084546A1 true WO2017084546A1 (zh) 2017-05-26

Family

ID=58718010

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/105720 WO2017084546A1 (zh) 2015-11-17 2016-11-14 基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备

Country Status (2)

Country Link
US (1) US10163528B2 (zh)
WO (1) WO2017084546A1 (zh)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10528714B2 (en) * 2017-01-11 2020-01-07 Samsung Electronics Co., Ltd. Method and apparatus for authenticating user using electrocardiogram signal
JP6765657B2 (ja) * 2017-06-27 2020-10-07 公立大学法人会津大学 認証装置、認証方法及びコンピュータプログラム
US11138770B2 (en) * 2017-11-06 2021-10-05 Shanghai United Imaging Healthcare Co., Ltd. Systems and methods for medical imaging
CN111772617B (zh) * 2019-04-04 2023-08-04 成都心吉康科技有限公司 Ecg显示滤波方法、装置及存储介质
KR20220073209A (ko) * 2020-11-26 2022-06-03 조선대학교산학협력단 심전도 신호로부터 설명 가능한 히트맵 영상에 근거한 질병을 판별하는 시스템 및 판별 방법
CN113598784B (zh) * 2021-08-25 2024-04-09 济南汇医融工科技有限公司 心律失常检测方法及系统

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101773394A (zh) * 2010-01-06 2010-07-14 中国航天员科研训练中心 身份识别方法及应用该方法的身份识别系统
CN103093133A (zh) * 2013-01-08 2013-05-08 西安电子科技大学 面向ieee802.15.6的生物身份认证方法
CN103584852A (zh) * 2012-08-15 2014-02-19 深圳中科强华科技有限公司 个性化心电图智能辅助诊断装置与方法
CN103610457A (zh) * 2013-11-19 2014-03-05 深圳先进技术研究院 一种心电信号的处理方法及系统
CN104055499A (zh) * 2014-06-16 2014-09-24 朱宇东 连续监控人体生理体征的可穿戴式智能手环及方法
CN104382590A (zh) * 2014-12-11 2015-03-04 赖大坤 一种结合心电时频域特征分析的可电击复律心律自动识别和归类方法
CN104382584A (zh) * 2014-11-24 2015-03-04 北京联合大学 一种可穿戴式心电信号监测评估系统及实现方法
CN104510463A (zh) * 2014-12-08 2015-04-15 华南理工大学 基于可穿戴装置的心电检测装置
CN104573458A (zh) * 2014-12-30 2015-04-29 深圳先进技术研究院 一种基于心电信号的身份识别方法、装置及系统
CN104720846A (zh) * 2015-02-04 2015-06-24 四川长虹电器股份有限公司 一种心脏健康评估方法
CN105286853A (zh) * 2015-11-17 2016-02-03 安徽华米信息科技有限公司 基于可穿戴设备的疾病检测方法及装置、可穿戴设备
CN105468951A (zh) * 2015-11-17 2016-04-06 安徽华米信息科技有限公司 通过心电特征进行身份识别的方法及装置、可穿戴设备

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9492084B2 (en) 2004-06-18 2016-11-15 Adidas Ag Systems and methods for monitoring subjects in potential physiological distress
US20090070266A1 (en) * 2007-09-07 2009-03-12 Shah Rahul C System and method for physiological data authentication and bundling with delayed binding of individual identification
US8200321B2 (en) 2009-05-20 2012-06-12 Sotera Wireless, Inc. Method for measuring patient posture and vital signs
EP2534597B1 (en) 2010-03-15 2018-10-17 Singapore Health Services Pte Ltd Method of predicting the survivability of a patient
US20130184517A1 (en) 2011-07-07 2013-07-18 Ronda Collier System and Method for Measuring and Controlling Stress
US8897863B2 (en) * 2012-05-03 2014-11-25 University Of Washington Through Its Center For Commercialization Arrhythmia detection using hidden regularity to improve specificity
EP3079571A4 (en) 2013-12-12 2017-08-02 Alivecor, Inc. Methods and systems for arrhythmia tracking and scoring
US9545227B2 (en) 2013-12-13 2017-01-17 Vital Connect, Inc. Sleep apnea syndrome (SAS) screening using wearable devices

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101773394A (zh) * 2010-01-06 2010-07-14 中国航天员科研训练中心 身份识别方法及应用该方法的身份识别系统
CN103584852A (zh) * 2012-08-15 2014-02-19 深圳中科强华科技有限公司 个性化心电图智能辅助诊断装置与方法
CN103093133A (zh) * 2013-01-08 2013-05-08 西安电子科技大学 面向ieee802.15.6的生物身份认证方法
CN103610457A (zh) * 2013-11-19 2014-03-05 深圳先进技术研究院 一种心电信号的处理方法及系统
CN104055499A (zh) * 2014-06-16 2014-09-24 朱宇东 连续监控人体生理体征的可穿戴式智能手环及方法
CN104382584A (zh) * 2014-11-24 2015-03-04 北京联合大学 一种可穿戴式心电信号监测评估系统及实现方法
CN104510463A (zh) * 2014-12-08 2015-04-15 华南理工大学 基于可穿戴装置的心电检测装置
CN104382590A (zh) * 2014-12-11 2015-03-04 赖大坤 一种结合心电时频域特征分析的可电击复律心律自动识别和归类方法
CN104573458A (zh) * 2014-12-30 2015-04-29 深圳先进技术研究院 一种基于心电信号的身份识别方法、装置及系统
CN104720846A (zh) * 2015-02-04 2015-06-24 四川长虹电器股份有限公司 一种心脏健康评估方法
CN105286853A (zh) * 2015-11-17 2016-02-03 安徽华米信息科技有限公司 基于可穿戴设备的疾病检测方法及装置、可穿戴设备
CN105468951A (zh) * 2015-11-17 2016-04-06 安徽华米信息科技有限公司 通过心电特征进行身份识别的方法及装置、可穿戴设备

Also Published As

Publication number Publication date
US10163528B2 (en) 2018-12-25
US20170242974A1 (en) 2017-08-24

Similar Documents

Publication Publication Date Title
WO2017084546A1 (zh) 基于可穿戴设备的用户关注信息确定方法、装置和可穿戴设备
CN105468951B (zh) 通过心电特征进行身份识别的方法及装置、可穿戴设备
CN105286853B (zh) 基于可穿戴设备的疾病检测方法及装置、可穿戴设备
US9946942B2 (en) Method, apparatus and system for biometric identification
US10572637B2 (en) User authentication method and apparatus based on electrocardiogram (ECG) signal
US10342447B2 (en) Systems and methods for model-based non-contact physiological data acquisition
Sadhukhan et al. R-peak detection algorithm for ECG using double difference and RR interval processing
CN111091116B (zh) 一种用于判断心律失常的信号处理方法及系统
US20160183812A1 (en) Biometric authentication method and apparatus
Abdeldayem et al. ECG-based human authentication using high-level spectro-temporal signal features
Shdefat et al. Utilizing ECG waveform features as new biometric authentication method
KR20160024600A (ko) 심전도 신호를 이용한 사용자 인증 방법 및 장치
KR20180072290A (ko) 개인 식별을 위한 심전도 획득 방법 및 그 심전도를 이용한 개인 식별 방법
Chin et al. Bayesian real-time QRS complex detector for healthcare system
Sadhukhan et al. Detection of ECG characteristic features using slope thresholding and relative magnitude comparison
CN110840443A (zh) 心电信号处理方法、心电信号处理装置和电子设备
KR20150068694A (ko) 사용자 인증 방법과 이를 수행하는 사용자 인증 장치 및 사용자 인증 시스템
Bastos et al. Double authentication model based on ppg and ecg signals
CN109875550B (zh) 一种心室除极关键点检测方法
CN114027804A (zh) 一种脉象诊断的方法、装置及可读存储介质
Canento et al. Review and comparison of real time electrocardiogram segmentation algorithms for biometric applications
JP2004248714A (ja) 生体信号を用いた認証方法および認証装置
Vuksanovic et al. Analysis of human electrocardiogram for biometric recognition using analytic and ar modeling extracted parameters
JP2023035888A (ja) 無線レーダー信号に基づいて心拍データを抽出するための装置及び方法
JP2015213714A (ja) 心電波形検出装置、心電波形検出方法、心電波形検出プログラム、及び撮像装置

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: 16865723

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16865723

Country of ref document: EP

Kind code of ref document: A1