Wearable electrocardiogram monitoring method and system

Abstract

The invention relates to a wearable electrocardiogram monitoring method and system, and the method comprises the steps: obtaining an electrocardiogram time-frequency diagram through electrocardiogram data collected by wearable electrocardiogram monitoring equipment, dividing the time-frequency diagram into a plurality of regions with the same size, initializing a climbing point based on a preset frequency band and the regions, and initializing a measurement matrix; the climbing point moves horizontally, after the climbing point moves horizontally, if the climbing point is located on the upper boundary and the lower boundary of the time-frequency diagram, the non-moving probability and the moving probability in the opposite direction of the boundary are calculated, and the vertical moving mode is determined according to the probabilities; a metric value is added to the climber in the metric matrix after the climber moves or does not move, and the metric value is the moving probability or the non-moving probability; otherwise, determining a vertical movement mode and a measurement matrix added value based on the upper and lower neighborhoods of the climbing point; and stopping moving when the number of times of movement reaches a preset value, obtaining a ridge line according to the measurement matrix, and judging the electrocardio condition by using the ridge line.

Description

Technical Field

[0001] The present invention relates to the field of electrocardiogram monitoring, and in particular to a wearable electrocardiogram monitoring method and system. Background Art

[0002] With the aging of the population, cardiovascular disease has become one of the major diseases threatening human health worldwide. Early warning and diagnosis of cardiovascular disease can help reduce mortality and improve the quality of life of patients. However, traditional ECG monitoring methods mainly rely on the equipment of medical institutions (such as static ECG machines or dynamic Holter devices), which need to be completed in hospitals. Not only are they limited by the distribution of medical resources, but they are also difficult to meet the needs of long-term and continuous health monitoring. At the same time, the rapid development of wearable devices has provided a new opportunity for individualized health management, especially in terms of portability and real-time performance. Wearable ECG monitoring devices collect ECG signals through highly sensitive sensors and combine them with advanced signal processing algorithms to provide ordinary users with low-cost and efficient health monitoring solutions. This technology can not only achieve early detection of diseases such as arrhythmias and myocardial ischemia, but also provide patients with long-term health data analysis to help doctors develop more accurate treatment plans.

[0003] However, due to the complex wearing environment, the ECG signals collected by wearable devices are usually interfered by noise (such as motion artifacts, environmental noise, and poor electrode contact), and the signal quality is low. Traditional analysis methods may show great limitations when processing these non-stationary signals. The method based on ridge extraction of time-frequency graphs can effectively capture the dynamic frequency components and local energy distribution characteristics in ECG signals, providing reliable feature data support for real-time analysis and anomaly detection. Wearable devices can accurately monitor indicators such as heart rate variability (HRV) and arrhythmias, thereby significantly improving the quality and efficiency of health management. However, how to accurately and quickly extract the ridges of ECG signals collected by wearable ECG monitoring devices is an urgent problem to be solved in this field. Summary of the invention

[0004] In order to improve the accuracy and speed of monitoring by a wearable ECG monitoring device, a wearable ECG monitoring method is provided in a first aspect of the present invention, the method comprising the following steps: The ECG data collected by the wearable ECG monitoring device is used to obtain the ECG time-frequency diagram, which is divided into multiple regions of the same size. The climbing points are initialized based on the preset frequency bands and regions, and the metric matrix is ​​initialized. The climbing point is first moved horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​increased as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; When the number of movements reaches a preset value, the movement stops, the ridge line is obtained according to the measurement matrix, and the ridge line is used to judge the ECG condition.

[0005] Optionally, the initializing the climbing point based on the preset frequency band and area is specifically: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

[0006] Optionally, the initializing a climbing point in each sub-area is specifically as follows: Taking the center point of each sub-region as the center, the probability distribution of the sub-region is obtained by using the distance decay function or the two-dimensional normal distribution; The product of the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram is used to obtain the weighted intensity value of the sub-region; The point with the largest weighted strength value in the sub-region is used as the location of the initial climbing point in the sub-region.

[0007] Optionally, the climbing point is first moved horizontally, specifically: The QRS wave group of the electrocardiogram data is detected to obtain the middle position of the QRS wave group. If the climbing point is not located at the left or right boundary of the time-frequency diagram, the distance between the climbing point and the middle position of the two nearest QRS wave groups is obtained, and the probability of the climbing point moving horizontally to the left and right is determined according to the distance. If the climbing point is located at the left or right boundary of the time-frequency diagram, it is determined to remain unchanged or move in the opposite direction of the boundary where the climbing point is located according to the same probability.

[0008] Optionally, the vertical movement mode and the metric matrix increase value are determined based on the upper and lower neighborhoods of the climbing point, specifically: Move downward or upward with the same probability. If the value of the climbing point after the move is greater than the value before the move, move it, and after the move is completed, add a metric value of 0.5 to the new position of the climbing point in the metric matrix; if the value of the climbing point after the move is not greater than the value before the move, determine the probabilities of moving up, moving down, and not moving according to the upper and lower neighbors of the climbing point before the vertical movement, move the climbing point according to the probabilities, and after the move is completed, add a metric value to the new position of the climbing point in the metric matrix as the probability corresponding to the movement mode.

[0009] Optionally, the probabilities of moving up, moving down, and not moving are determined according to the upper and lower neighborhoods of the climbing point before the vertical movement, specifically: Compare the values ​​of up, down, and the current position, calculate the proportion of each in the total, and use these proportions as the corresponding move probabilities.

[0010] In a second aspect of the present invention, a wearable electrocardiogram monitoring system is provided, the system comprising the following modules: An initialization module is used to obtain a time-frequency diagram of the ECG from the ECG data collected by the wearable ECG monitoring device, divide the time-frequency diagram into a plurality of regions of the same size, initialize climbing points based on preset frequency bands and regions, and initialize a metric matrix; A moving module is used to first move the climbing point horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​added as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; The monitoring module is used to stop moving when the number of movements reaches a preset value, obtain a ridge line according to the measurement matrix, and use the ridge line to judge the ECG condition.

[0011] Optionally, the initializing the climbing point based on the preset frequency band and area is specifically: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

[0012] Optionally, the initializing a climbing point in each sub-area is specifically as follows: Taking the center point of each sub-region as the center, the probability distribution of the sub-region is obtained by using the distance decay function or the two-dimensional normal distribution; The product of the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram is used to obtain the weighted intensity value of the sub-region; The point with the largest weighted strength value in the sub-region is used as the location of the initial climbing point in the sub-region.

[0013] Optionally, the climbing point is first moved horizontally, specifically: The QRS wave group of the electrocardiogram data is detected to obtain the middle position of the QRS wave group. If the climbing point is not located at the left or right boundary of the time-frequency diagram, the distance between the climbing point and the middle position of the two nearest QRS wave groups is obtained, and the probability of the climbing point moving horizontally to the left and right is determined according to the distance. If the climbing point is located at the left or right boundary of the time-frequency diagram, it is determined to remain unchanged or move in the opposite direction of the boundary where the climbing point is located according to the same probability.

[0014] Optionally, the vertical movement mode and the metric matrix increase value are determined based on the upper and lower neighborhoods of the climbing point, specifically: Move downward or upward with the same probability. If the value of the climbing point after the move is greater than the value before the move, move it, and after the move is completed, add a metric value of 0.5 to the new position of the climbing point in the metric matrix; if the value of the climbing point after the move is not greater than the value before the move, determine the probabilities of moving up, moving down, and not moving according to the upper and lower neighbors of the climbing point before the vertical movement, move the climbing point according to the probabilities, and after the move is completed, add a metric value to the new position of the climbing point in the metric matrix as the probability corresponding to the movement mode.

[0015] Optionally, the probabilities of moving up, moving down, and not moving are determined according to the upper and lower neighborhoods of the climbing point before the vertical movement, specifically: Compare the values ​​of up, down, and the current position, calculate the proportion of each in the total, and use these proportions as the corresponding move probabilities.

[0016] In addition, the present invention also provides a computer program, which implements the above method when executed by a processor.

[0017] The present invention initializes the climbing point based on the preset frequency band to avoid meaningless climbing in unimportant frequency bands; and determines the vertical movement mode according to the upper and lower neighborhoods of the climbing point, so that the climbing process can adapt to the local characteristics of the time-frequency graph, more effectively find the area with higher energy, and avoid the climbing point oscillating near the local minimum or noise point. In addition, the movement probability or non-movement probability is used as the added value of the metric matrix, so that the metric matrix can reflect the randomness and uncertainty of the climbing process, while also taking into account the exploration and convergence of the climbing. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of Embodiment 1; Figure 2 This is a schematic diagram of an electrocardiogram; Figure 3 It is a schematic diagram of the QRS complex; Figure 4 This is a structural diagram of the second embodiment. DETAILED DESCRIPTION

[0019] In this article, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In a first embodiment of the present invention, Figure 1 As shown, a wearable electrocardiogram monitoring method is provided, the method comprising the following steps: S1, obtaining a time-frequency graph of the ECG from ECG data collected by a wearable ECG monitoring device, dividing the time-frequency graph into multiple regions of the same size, initializing climbing points based on preset frequency bands and regions, and initializing a metric matrix; The ECG signal is collected through wearable devices, where the wearable devices include but are not limited to ECG patches, smart bracelets, etc. Figure 2An electrocardiogram of electrocardiogram data is shown. After the electrocardiogram data is collected, the electrocardiogram signal is converted into a time-frequency diagram to obtain a time-frequency diagram of the electrocardiogram signal, wherein the horizontal axis of the time-frequency diagram is time, the vertical axis is frequency, and the value of the time-frequency diagram represents intensity. After the time-frequency diagram is generated, the time-frequency diagram is divided into a plurality of regions of equal size, and a plurality of climbing points (Climbers) are initialized according to a preset frequency range, i.e., a preset frequency band, and the regions. In a specific embodiment, the preset frequency is 0.5-4Hz or 4-15Hz; in another embodiment, the preset frequency band is determined according to the user's own condition wearing the electrocardiogram monitoring device, for example, if there is atrial fibrillation, the high frequency is set to the preset frequency band, etc.

[0022] In order to extract the ridge line faster, the preset frequency band and the region are used to initialize the climbing point. Specifically, if the region is located in or partially located in the preset frequency band, the region is divided into more fine-grained areas to obtain sub-regions, and a climbing point is initialized in each sub-region. If the region and the preset frequency band have no intersection, fine-grained division is not performed, and a climbing point is initialized in the region.

[0023] In yet another embodiment, the initializing the climbing point based on the preset frequency band and area is specifically: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

[0024] The preset frequency band is a range value. In the time-frequency diagram, a preset frequency band is a region. If the region overlaps with the preset frequency band, that is, there is an intersection, but the area of ​​the overlapping part is less than half of the area of ​​the rectangular region itself, then the rectangular region is vertically divided into two smaller rectangular regions, that is, two sub-regions. If the overlapping area between the region and the preset frequency band is greater than half of the area of ​​the region, then the region is evenly divided into four smaller regions. If there is no overlap between the region and the preset frequency band, then the region is kept unchanged and the region is treated as a separate sub-region. After the sub-regions are divided, a climbing point is set in each sub-region. The climbing point is the starting search point in the Crazy Climber algorithm, which looks for the path with the highest energy, that is, the ridge line. In one embodiment, the center position of each sub-region is used as the position of the initial climbing point. For example, if the size of a sub-region is 5×5, the starting position of the climbing point is the center of the sub-region.

[0025] In another embodiment, the probability distribution of each sub-region is obtained by using the distance decay function or two-dimensional normal distribution with the center point of each sub-region as the center; the weighted intensity value of the sub-region is obtained by multiplying the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram; the point with the largest weighted intensity value in the sub-region is used as the position of the initial climbing point in the sub-region. Specifically, with the center point as the reference, a probability distribution diagram is generated to indicate the possibility of different positions in the sub-region as the starting position of the climbing point. There are many ways to generate a probability distribution diagram, preferably using a distance decay function or a two-dimensional normal distribution, wherein the distance function satisfies that the closer the point is to the center point, the higher the probability; the farther the point is from the center point, the lower the probability; for example, a Gaussian function or other attenuation functions can be used; the two-dimensional normal distribution takes the center point as the mean, and creates a two-dimensional normal distribution in the sub-region, with the center point having the highest probability, and the probability is lower as it goes outward, forming a probability peak centered on the center point.

[0026] The calculated probability distribution and the intensity distribution of the time-frequency graph are multiplied point by point, so that the selection of the starting point is combined with the actual energy of the signal. After calculating the weighted intensity values ​​of all points in the sub-region, the point with the maximum weighted intensity value is selected as the starting position of the climbing point of the sub-region. This point has both a higher probability of being a starting point and a higher signal energy, so it is more likely to be located on the ridge line. If a point has a high probability even if it is very close to the center, but if the energy of the point on the time-frequency graph is very low, then its weighted intensity value will also be low, thereby reducing the possibility of it being selected as a climbing point. This not only prevents the climber from being located at the boundary of the sub-region, causing the climbers of the two adjacent sub-regions to be too close, but also allows the climber to be initialized to a position that is close to the center of the sub-region and has a larger intensity value, combined with the intensity value of the time-frequency graph.

[0027] After initializing the climber, initialize the metric matrix. The size of the metric matrix is ​​the same as the size of the time-frequency graph, and the element values ​​in the initialized metric matrix are all 0.

[0028] S2, the climbing point is first moved horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​increased as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; When the climber climbs or moves on the time-frequency graph, the present invention is the same as the Crazy Climber algorithm, first moving horizontally and then moving vertically. In one embodiment, the horizontal movement adopts the same horizontal movement mode as the Crazy Climber.

[0029] The QRS complex is one of the most prominent features of the ECG signal. Figure 3 The QRS complex is shown. In yet another embodiment, the climbing point is first moved horizontally, specifically: The QRS wave group of the electrocardiogram data is detected to obtain the middle position of the QRS wave group. If the climbing point is not located at the left or right boundary of the time-frequency diagram, the distance between the climbing point and the middle position of the two nearest QRS wave groups is obtained, and the probability of the climbing point moving horizontally to the left and right is determined according to the distance. If the climbing point is located at the left or right boundary of the time-frequency diagram, it is determined to remain unchanged or move in the opposite direction of the boundary where the climbing point is located according to the same probability.

[0030] The QRS wave group is detected on the original ECG data to determine the center position of each QRS wave group on the time axis of the ECG. Since the time axis of the ECG and the time axis of the time-frequency diagram can correspond, the middle position of the QRS wave group on the time-frequency diagram is determined. The middle position is the middle position between the starting point and the ending point of the QRS wave group. For example, if the starting time of a QRS wave group is 3s and the ending time is 9s, the middle position is 6s.

[0031] If the climbing point is not located at the left or right boundary of the time-frequency diagram, obtain the middle position of the two QRS wave groups closest to the climber, and calculate the distance from the climbing point to the center position of the two QRS wave groups. Since the horizontal coordinate of the climbing point on the time-frequency diagram corresponds to time, the distance is also the time difference. According to the calculated distance, determine the probability of the climbing point moving horizontally to the left and right. In one embodiment, the probability of moving to a certain direction is inversely proportional to the distance to the nearest QRS wave group in that direction. The closer the distance, the greater the probability. More specifically, an inverse proportional function is used to determine the probability of moving to the left and right. For example, if the ratio of the distance of a climber from the nearest center position to the left to the nearest center position to the right is 1:3, then the probability of moving to the left is 0.75, and the probability of moving to the right is 0.25. If the climbing point is located at the leftmost or rightmost boundary of the time-frequency diagram, with the same probability, for example 0.5, choose to keep the current position or move in the opposite direction of the boundary.

[0032] After the horizontal movement, the vertical movement is performed. If the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value is added to the climber in the metric matrix as the probability of moving or not moving.

[0033] Determine whether the current climbing point is located at the top or bottom boundary of the time-frequency graph. If the climbing point is located at the boundary, calculate the probability of not moving, that is, the probability of keeping the current vertical position unchanged; and the probability of moving in the opposite direction, that is, the probability of moving in the opposite direction of the boundary. The sum of these two probabilities is usually 1, for example, they can be set to 0.5 each, or they can be adjusted according to the specific application scenario. According to the calculated probability, randomly decide whether the climbing point remains stationary or moves in the opposite direction. If the climbing point moves, the value added to the corresponding position in the metric matrix is ​​the probability of moving in the opposite direction. If the climbing point does not move, the value added to the corresponding position in the metric matrix is ​​the probability of not moving. Among them, the corresponding position of the metric matrix is ​​the position of the climber's coordinates in the metric matrix after the value moves vertically or does not move.

[0034] Assume that the climbing point is located at the top boundary of the time-frequency graph, with a probability of not moving of 0.4 and a probability of moving downward of 0.6. If the climbing point is randomly determined to move downward, the corresponding position in the metric matrix needs to be increased by 0.6. If the climbing point is randomly determined to remain stationary, the corresponding position in the metric matrix needs to be increased by 0.4.

[0035] If the climbing point is not located at the boundary, that is, the climbing point is not on the upper or lower boundary, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point. In one embodiment, the upper and lower neighborhoods are the areas directly above and below the climbing point. The energy intensities of the adjacent positions above and below the climbing point are compared, and the direction with greater energy is selected for movement; or, the probability of moving upward and downward is calculated based on the energy intensity ratio of the upper and lower neighborhoods, and then the moving direction is randomly selected based on the probability. For example, if the energy of the upper neighborhood is twice that of the lower neighborhood, the probability of moving upward can be 2 / 3, and the probability of moving downward can be 1 / 3. After the movement, the metric matrix is ​​updated. If the movement mode is determined by selecting the neighborhood with the largest energy, the added value can be determined based on the energy difference between the upper and lower neighborhoods; if the probabilistic movement mode is adopted, the added value is determined based on the probability of movement. For example, the energy of the neighborhood above the climbing point is 10, and the energy of the neighborhood below is 5. If the neighborhood with the largest energy is selected for movement, the climbing point moves upward, and a value related to the energy difference 5 can be added to the metric matrix, for example, a proportional value positively correlated with the energy difference is added, and the proportional value is limited between 0 and 1. In one embodiment, the upper neighborhood is a neighborhood above the climber after horizontal movement, and the lower neighborhood is a neighborhood below the climber after horizontal movement, that is, the upper and lower neighborhoods are both a point. In another embodiment, the upper neighborhood and the lower neighborhood are the intensity averages of an area, the area of ​​the upper neighborhood is a preset size area with the upper neighborhood point as the center after the climber moves horizontally, and the area of ​​the lower neighborhood is a preset size area with the lower neighborhood point as the center after the climbing point moves horizontally.

[0036] In yet another embodiment, the vertical movement mode and the metric matrix increase value are determined based on the upper and lower neighborhoods of the climbing point, specifically: Move downward or upward with the same probability. If the value of the climbing point after the move is greater than the value before the move, move it, and after the move is completed, add a metric value of 0.5 to the new position of the climbing point in the metric matrix; if the value of the climbing point after the move is not greater than the value before the move, determine the probabilities of moving up, moving down, and not moving according to the upper and lower neighbors of the climbing point before the vertical movement, move the climbing point according to the probabilities, and after the move is completed, add a metric value to the new position of the climbing point in the metric matrix as the probability corresponding to the movement mode.

[0037] The climbing point will try to move up or down one step with equal probability, for example, 50% each. After the move, the value of the new position (i.e., the upper or lower adjacent position) on the time-frequency graph will be compared with the value before the move. If the value after the move is greater than the value before the move, it is confirmed that the move is valid and the climbing point will move to the new position. A metric value of 0.5 is added to the new position of the climbing point in the metric matrix. If the value after the move is not greater than the value before the move, that is, it has moved to a position with lower or equal energy, the probability of moving up, moving down, and staying still is determined according to the values ​​of the upper and lower neighbors of the climbing point before the move. Specifically, the values ​​of the upper, lower, and current positions are compared, the proportion of each in the total is calculated, and these proportions are used as the corresponding movement probabilities. For example, if the value of the upper neighbor is 3, the value of the lower neighbor is 2, and the value of the current position is 1, then the probability of moving up is 3 / (3+2+1)=0.5, the probability of moving down is 2 / (3+2+1)=0.33, and the probability of staying still is 1 / (3+2+1)=17. According to the calculated probability, randomly choose to move up, move down, or stay still. After the movement (or no movement) is completed, the metric value that needs to be added to the new position of the climbing point in the metric matrix is ​​the probability corresponding to this movement method.

[0038] S3, when the number of movements reaches a preset value, the movement is stopped, a ridge line is obtained according to the measurement matrix, and the ridge line is used to judge the ECG condition.

[0039] After the iteration is stopped, the metric matrix is ​​obtained, and the metric matrix is ​​recursively transformed in the time direction to obtain the ridge line, and the ridge line is filtered according to the length of the ridge line to finally obtain the extracted ridge line. The ECG signal is analyzed according to the frequency position of the ridge line on the time-frequency diagram, the shape change of the ridge line, and the number of ridge lines. In one embodiment, the features of the ridge line are input into a support vector machine or the like to identify the ECG state.

[0040] In a first embodiment of the present invention, a wearable electrocardiogram monitoring system is provided. Figure 4 As shown, the system includes the following modules: An initialization module is used to obtain a time-frequency diagram of the ECG from the ECG data collected by the wearable ECG monitoring device, divide the time-frequency diagram into a plurality of regions of the same size, initialize climbing points based on preset frequency bands and regions, and initialize a metric matrix; A moving module is used to first move the climbing point horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​added as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; The monitoring module is used to stop moving when the number of movements reaches a preset value, obtain a ridge line according to the measurement matrix, and use the ridge line to judge the ECG condition.

[0041] Optionally, the initializing the climbing point based on the preset frequency band and area is specifically: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

[0042] Optionally, the initializing a climbing point in each sub-area is specifically as follows: Taking the center point of each sub-region as the center, the probability distribution of the sub-region is obtained by using the distance decay function or the two-dimensional normal distribution; The product of the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram is used to obtain the weighted intensity value of the sub-region; The point with the largest weighted strength value in the sub-region is used as the location of the initial climbing point in the sub-region.

[0043] Optionally, the climbing point is first moved horizontally, specifically: The QRS wave group of the electrocardiogram data is detected to obtain the middle position of the QRS wave group. If the climbing point is not located at the left or right boundary of the time-frequency diagram, the distance between the climbing point and the middle position of the two nearest QRS wave groups is obtained, and the probability of the climbing point moving horizontally to the left and right is determined according to the distance. If the climbing point is located at the left or right boundary of the time-frequency diagram, it is determined to remain unchanged or move in the opposite direction of the boundary where the climbing point is located according to the same probability.

[0044] Optionally, the vertical movement mode and the metric matrix increase value are determined based on the upper and lower neighborhoods of the climbing point, specifically: Move downward or upward with the same probability. If the value of the climbing point after the move is greater than the value before the move, move it, and after the move is completed, add a metric value of 0.5 to the new position of the climbing point in the metric matrix; if the value of the climbing point after the move is not greater than the value before the move, determine the probabilities of moving up, moving down, and not moving according to the upper and lower neighbors of the climbing point before the vertical movement, move the climbing point according to the probabilities, and after the move is completed, add a metric value to the new position of the climbing point in the metric matrix as the probability corresponding to the movement mode.

[0045] Optionally, the probabilities of moving up, moving down, and not moving are determined according to the upper and lower neighborhoods of the climbing point before the vertical movement, specifically: Compare the values ​​of up, down, and the current position, calculate the proportion of each in the total, and use these proportions as the corresponding move probabilities.

[0046] Through the description of the above implementation methods, technicians in this field can clearly understand that each implementation method can be implemented by adding a necessary general hardware platform, and of course can also be implemented by combining hardware and software. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a computer product, and the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it, and other embodiments may also be used. Although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wearable electrocardiogram monitoring method, characterized in that: The method comprises the following steps: The ECG data collected by the wearable ECG monitoring device is used to obtain the ECG time-frequency diagram, which is divided into multiple regions of the same size. The climbing points are initialized based on the preset frequency bands and regions, and the metric matrix is ​​initialized. The climbing point is first moved horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​increased as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; When the number of movements reaches a preset value, the movement stops, the ridge line is obtained according to the measurement matrix, and the ridge line is used to judge the ECG condition.

2. The method according to claim 1, characterized in that The initialization of climbing points based on the preset frequency band and area is specifically as follows: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

3. The method according to claim 2, characterized in that Initializing a climbing point in each sub-area is as follows: Taking the center point of each sub-region as the center, the probability distribution of the sub-region is obtained by using the distance decay function or the two-dimensional normal distribution; The product of the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram is used to obtain the weighted intensity value of the sub-region; The point with the largest weighted strength value in the sub-region is used as the location of the initial climbing point in the sub-region.

4. The method according to claim 1, characterized in that The climbing point is first moved horizontally, specifically: The QRS wave group of the electrocardiogram data is detected to obtain the middle position of the QRS wave group. If the climbing point is not located at the left or right boundary of the time-frequency diagram, the distance between the climbing point and the middle position of the two nearest QRS wave groups is obtained, and the probability of the climbing point moving horizontally to the left and right is determined according to the distance. If the climbing point is located at the left or right boundary of the time-frequency diagram, it is determined to remain unchanged or move in the opposite direction of the boundary where the climbing point is located according to the same probability.

5. The method according to claim 1, characterized in that The vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point, specifically: Move downward or upward with the same probability. If the value of the climbing point after the move is greater than the value before the move, move it, and after the move is completed, add a metric value of 0.5 to the new position of the climbing point in the metric matrix; if the value of the climbing point after the move is not greater than the value before the move, determine the probabilities of moving up, moving down, and not moving according to the upper and lower neighbors of the climbing point before the vertical movement, move the climbing point according to the probabilities, and after the move is completed, add a metric value to the new position of the climbing point in the metric matrix as the probability corresponding to the movement mode.

6. The method according to claim 5, characterized in that The probability of moving up, moving down, and not moving is determined according to the upper and lower neighborhoods of the climbing point before vertical movement, specifically: Compare the values ​​of up, down, and the current position, calculate the proportion of each in the total, and use these proportions as the corresponding move probabilities.

7. A wearable electrocardiogram monitoring system, characterized in that: The system includes the following modules: An initialization module is used to obtain a time-frequency diagram of the ECG from the ECG data collected by the wearable ECG monitoring device, divide the time-frequency diagram into a plurality of regions of the same size, initialize climbing points based on preset frequency bands and regions, and initialize a metric matrix; A moving module is used to first move the climbing point horizontally. After the horizontal movement, if the climbing point is located at the upper and lower boundaries of the time-frequency graph, the probability of not moving and the probability of moving in the opposite direction of the boundary are calculated, and the vertical movement mode is determined according to the probabilities. After moving or not moving, the metric value at the climber in the metric matrix is ​​added as the probability of moving or not moving; otherwise, the vertical movement mode and the added value of the metric matrix are determined based on the upper and lower neighborhoods of the climbing point; The monitoring module is used to stop moving when the number of movements reaches a preset value, obtain a ridge line according to the measurement matrix, and use the ridge line to judge the ECG condition.

8. The system according to claim 7, characterized in that The initialization of climbing points based on the preset frequency band and area is specifically as follows: Obtain a preset frequency band of the time-frequency graph; if the region and the preset frequency band have an intersection and the area of ​​the intersection is less than half of the region, divide the region into two sub-regions on average; if the region and the preset frequency band have an intersection and the area of ​​the intersection is greater than half of the region, divide the region into four sub-regions on average; if the region and the preset frequency band have no intersection, treat the region as a sub-region; initialize a climbing point in each sub-region.

9. The system according to claim 8, characterized in that Initializing a climbing point in each sub-area is as follows: Taking the center point of each sub-region as the center, the probability distribution of the sub-region is obtained by using the distance decay function or the two-dimensional normal distribution; The product of the probability distribution of the sub-region and the intensity distribution of the sub-region in the time-frequency diagram is used to obtain the weighted intensity value of the sub-region; The point with the largest weighted strength value in the sub-region is used as the location of the initial climbing point in the sub-region.

10. A computer program, characterized in that When the computer program is executed by a processor, the computer program implements the method according to any one of claims 1 to 6.

Images