A non-invasive ECG signal auxiliary monitoring and alarm system and method

Abstract

The application discloses a non-invasive ECG signal auxiliary monitoring and alarm system and method, relates to the fields of electrocardiogram signal automatic analysis and diagnosis and digital image processing, and comprises a video stream acquisition module, a picture interception module, a denoising and binarization module, a key point monitoring module, a labeled video stream module and a video stream display module. The video stream acquisition module acquires a video stream from an electrophysiological recording system; the picture interception module intercepts signal pictures at preset positions in the video stream; the denoising and binarization module pre-processes the signal pictures to obtain clean signal pictures; the key point monitoring module detects preset key points in the signal pictures according to pixels to obtain detection signals of the key points; the labeled video stream module detects the key points and gives a prompt; and the video stream display module displays the signal pictures in the form of the video stream on a result display screen, so that a user can observe in real time.

Description

Technical Field

[0001] This invention relates to the fields of automatic analysis and diagnosis of electrocardiogram (ECG) signals and digital image processing, and particularly to a non-invasive ECG signal-assisted monitoring and alarm system and method. Background Technology

[0002] Cardiovascular disease is one of the leading causes of death worldwide, and cardiac surgeries (such as cardiac pacing and cardiac catheterization) play a crucial role in treating these diseases. During these procedures, real-time monitoring of electrocardiogram (ECG) signals is essential for surgical success and patient safety. Currently, ECG monitoring equipment is widely used in clinical practice to assist surgeons by analyzing cardiac electrical activity signals.

[0003] However, existing electrocardiogram (ECG) monitoring equipment still has certain limitations, including:

[0004] 1) Reliance on equipment manufacturer's signal interface: Most ECG monitoring devices directly process raw signals, using specific algorithms to achieve signal analysis and alarm functions. However, these devices are usually proprietary to the manufacturers, and customized needs require open interfaces or redesigned hardware, resulting in long development cycles and high costs. For example, during catheter positioning, doctors need to repeatedly adjust the catheter to capture the His bundle signal, and existing basic equipment often cannot provide sufficient signal guidance. In addition, some hospitals are still using relatively basic ECG monitors due to budget constraints or equipment upgrade issues. These devices lack advanced signal analysis and alarm functions, and can only provide basic ECG display, making it difficult to meet the precise monitoring needs of critical signals during surgery. For example, in cardiac pacing surgery, doctors need to monitor the relationship between ventricular electrodes and ECG signals in real time, and the functionality of existing equipment may not meet the customized needs of specific hospitals.

[0005] 2) Insufficient postoperative recording and analysis: Existing monitoring equipment mainly focuses on real-time alarms and usually cannot record and annotate signal changes during the operation in detail. For example, in complex percutaneous coronary intervention (PCI), dynamic changes in the ST segment can reflect the patient's myocardial blood supply status, but after the operation, the equipment may only save a brief record, which limits the doctor's ability to review and analyze the data after the operation, and is not conducive to summarizing experience or using it for medical teaching.

[0006] Therefore, those skilled in the art are dedicated to developing a new auxiliary monitoring and alarm system and method to overcome the aforementioned limitations in the existing technology. Summary of the Invention

[0007] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is how to get rid of the dependence on the signal interface of the equipment manufacturer and how to improve postoperative recording and analysis, so as to assist doctors in performing surgical operations by analyzing cardiac electrical activity signals.

[0008] To achieve the above objectives, the present invention provides a non-invasive ECG signal-assisted monitoring and alarm system, including a video stream acquisition module, an image capture module, a noise reduction and binarization module, a key point monitoring module, a labeled video stream module, and a video stream display module;

[0009] The video stream acquisition module acquires the video stream from the electrophysiological recording system;

[0010] The image cropping module is connected to the video stream acquisition module and extracts the signal image at a preset position from the acquired video stream;

[0011] The denoising and binarization module is connected to the image cropping module and preprocesses the cropped signal image, including removing noise points and binarizing the image, to obtain a clean signal image.

[0012] The key point monitoring module is connected to the denoising and binarization module, and detects the detection signals of pre-set key points by pixel in the pre-processed signal image.

[0013] The labeled video stream module is connected to the key point monitoring module. If the detection signal of the key point exceeds the preset threshold range, the corresponding position is marked with a bright color in the signal image and a prompt is given.

[0014] The video stream display module is connected to the labeled video stream module and displays the signal image on the result display screen in the form of a video stream, which is convenient for users to observe in real time.

[0015] Furthermore, the video stream acquisition module acquires the video stream from the electrophysiological recording system through a video capture card, connects it to the computing server via a USB interface, and acquires the video stream frame by frame using the VideoCapture class of the OpenCV library.

[0016] Furthermore, the video capture card is used to transmit the real-time images of the electrophysiological recording system frame by frame to the computing server. The video capture card used has acquisition parameters of 1080 / 60Hz, a data transmission interface of USB 3.0, and is backward compatible with USB 2.0.

[0017] Furthermore, the key points preset in the key point monitoring module are the starting point and peak point of each beat signal, and the peak time of each beat signal is calculated based on the starting point and the peak point.

[0018] Furthermore, the threshold range preset in the labeled video stream module is a set range regarding the peak time.

[0019] This invention also provides a non-invasive ECG signal-assisted monitoring and alarm method, the method comprising the following steps:

[0020] Step 1: Use a video capture card to acquire a real-time video stream of the entire screen content from the electrophysiological recording system;

[0021] Step 2: Extract the signal image from the pre-defined position in the acquired video stream;

[0022] Step 3: Preprocess the captured signal image, including removing noise points and binarizing the image, to obtain a clean signal image;

[0023] Step 4: For the preprocessed signal image, use the peak time detection algorithm to detect key points and determine if there are key points; if there are key points, mark the location of the key points and provide sound prompts.

[0024] Step 5: Display the signal image as a video stream on the result display screen for real-time observation by the user.

[0025] Furthermore, the preset position in step 2 includes y_offset, x_offset, height, and width, where y_offset refers to the offset of a pixel in height, x_offset refers to the offset of a pixel in width, y_offset and x_offset determine the starting point, height represents the height of the image to be cropped, and width represents the width of the image to be cropped. Height and width determine the range of the image to be cropped.

[0026] Furthermore, the signal image captured in step 2 is in RGB format, and the noise removal in step 3 includes setting thresholds in advance on the three different color channels of the signal image, removing pixels of different colors to filter out noise points, removing points not in the connected region, and using a connected component algorithm to extract the signal to be analyzed.

[0027] In step 3, the binarized image will also be used to calculate the position pos of the first white pixel in each column and the distance value from the first white pixel to the bottom of the signal image.

[0028] Furthermore, step 4 includes the following sub-steps:

[0029] Step 4.1: Based on the maximum value of the pixel detection area, obtain the peak position, including calculating the average value of the distance value, denoted as average, and using the find_peaks function in the scipy.signal module to find the local maximum value of the signal in the signal image;

[0030] Step 4.2: Remove redundant peaks, including removing peaks in the first 30 pixels and the last 12 pixels; remove small peaks, including removing peaks with an average left-right drop of less than 3.5, where the drop refers to the average difference between the peak value and the surrounding 12 pixels; remove isolated peaks, including traversing the peak list and adding peaks with a distance between adjacent peaks between (5, 100) to the peaks_add_list; return the peaks_add_list after deduplication and sorting, which contains the peak indices that meet the criteria.

[0031] Step 4.3, Double Peak Time Difference Detection: After calculating the peak positions in Step 4.1, the positions corresponding to each pair of peaks are obtained, denoted as pos_pair = [x10, x11, x20, x21, x30, x31...], where x10 and x11 are a pair of large and small peak positions, x20 and x21 are a pair of large and small peak positions, and so on; subtracting each pair yields the time interval between the two peaks, denoted as t = [t1, t2, t3...].

[0032] Step 4.4: Mark the double-peak time difference detection results, including marking the double-peak time difference on the signal image, marking nodes that exceed the threshold with a conspicuous color, and providing an audio prompt;

[0033] Step 4.5: Multi-peak separation detection, including dividing the peaks into groups and selecting the main peak in each group, recording it in the Not_Split library, and then determining whether there are split peaks, recording them in the Split library;

[0034] Step 4.6: Mark the multi-peak separation detection results, including marking all detected peaks on the image. If a split peak is detected, mark it with a conspicuous color and provide an audio prompt.

[0035] Furthermore, in step 4.5, the standard for grouping peaks is to group those within 100 pixels into one group; the standard for determining whether a peak is split is that the distance between two peaks is between 20 and 50 pixels, the numerical difference is less than 10 pixels, and the value of one of the peaks is greater than 140 pixels.

[0036] The non-invasive ECG signal-assisted monitoring and alarm system and method provided by this invention has at least the following technical effects:

[0037] 1. The technical solution provided by this invention does not require the use of original signal data, and has the advantages of strong flexibility and adaptability and reduced equipment dependence. In actual scenarios, since medical instruments (systems) do not have open interfaces, it is generally impossible to obtain original signal data. However, the method based on digital image processing can directly simulate visual monitoring of key points.

[0038] 2. The technical solution provided by this invention has excellent real-time performance. The detection of the four signals can be controlled within 0.5 seconds, which can be used for early warning during the operation.

[0039] 3. The data analyzed by the technical solution provided by this invention is at the pixel level, and the measurement error of peak time is almost zero, which can significantly improve the efficiency and accuracy of electrical signal monitoring in cardiac surgery.

[0040] 4. The technical solution provided by this invention does not require a large number of labeled materials for training, and is especially suitable for scenarios that require personalized monitoring or where old equipment cannot meet complex functions.

[0041] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0042] Figure 1 This is a physical connection diagram of a preferred embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the module flow of a preferred embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the algorithm logic of a preferred embodiment of the present invention. Detailed Implementation

[0045] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0046] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0047] This invention proposes a real-time electrocardiogram (ECG) signal monitoring and alarm system based on video streaming. The system directly acquires the signal waveforms displayed on the screen of the ECG monitoring device, simulating the process of a doctor observing an ECG through image processing technology. It does not rely on the manufacturer's open signal interface, achieving accurate capture, real-time alarm, and automatic annotation of key electrical signals. Typical application scenarios include: monitoring key signals during cardiac pacing surgery, including capturing specific signals from certain leads in real time to help doctors accurately locate electrode positions and reduce the risk of misoperation; auxiliary monitoring in complex interventional surgeries, including capturing subtle abnormalities (such as ST segment changes) in ECG waveforms during long, high-intensity surgeries to provide doctors with real-time decision support; and postoperative signal annotation and analysis, including automatically marking key intraoperative nodes to facilitate postoperative review and summarization by doctors, improving medical teaching effectiveness.

[0048] Specifically, the non-invasive ECG signal-assisted monitoring and alarm system provided in this embodiment of the invention is mainly an external signal detection and calculation module, which is equivalent to an external plug-in of the electrophysiological recording system. This module is used to monitor signal fluctuations on the display screen in real time, and then, based on medical knowledge, determines whether any points requiring surgical attention have occurred. In situations where real-time data cannot be directly obtained, this embodiment of the invention proposes an alternative solution—HDMI / VGA video capture processing.

[0049] Example 1

[0050] like Figure 1 and Figure 2 As shown in the illustration, a non-invasive ECG signal-assisted monitoring and alarm system provided by an embodiment of the present invention includes a video stream acquisition module, an image capture module, a noise reduction and binarization module, a key point monitoring module, an annotated video stream module, and a video stream display module. Since this system is an auxiliary tool for medical personnel during surgery, real-time performance and accuracy are crucial. Doctors expect to detect the key signal within one heartbeat's time after its appearance, i.e., 0.6s-1s; the worst-case scenario is that the key signal is detected before the image leaves the screen, i.e., within 3s. To achieve these requirements, the system's built-in algorithm is a key signal detection method based on digital image processing, specifically based on pixel-level abrupt changes in the image. The accuracy and real-time performance of this method meet the needs of doctors, making it highly suitable for practical applications.

[0051] Among them, the video stream acquisition module acquires the video stream from the electrophysiological recording system, that is, the real-time video stream of the hospital's electrocardiogram display screen is captured by the video stream splitting and capture device;

[0052] The image capture module connects to the video stream acquisition module and captures the signal image at a pre-defined position in the acquired video stream, because it is not necessary to process the pixels of the entire screen in subsequent analysis.

[0053] The denoising and binarization module is connected to the image cropping module to preprocess the cropped signal image, including removing noise points and binarizing the image to obtain a clean signal image.

[0054] The key point monitoring module is connected to the denoising and binarization module. It detects the detection signals of pre-set key points on a pixel-by-pixel basis for the pre-processed signal image. That is, it identifies key waveforms (such as R-wave, T-wave, etc.) and key signals through algorithms. The target point is accurately identified and the response time is in milliseconds.

[0055] The video stream annotation module is connected to the key point monitoring module. If the detection signal of a pre-set key point exceeds the pre-set threshold range, the corresponding position is marked with a bright color in the signal image and a prompt is given.

[0056] The video stream display module connects to the labeled video stream module, displaying the signal image on the result display screen in the form of a video stream, making it convenient for users to observe in real time.

[0057] Specifically, for electrophysiological recording systems, medical staff can select a walking speed of 100 to 200 during the operation to display the patient's heart rate, and each screen can display the number of pacings in the range of 3 to 6.

[0058] Specifically, the video stream acquisition module acquires the video stream from the electrophysiological recording system via a video capture card and connects it to the computing server via a USB interface. It uses the VideoCapture class from the OpenCV library to acquire the video stream frame by frame. The video capture card transmits the real-time images from the electrophysiological recording system to the computing server frame by frame. The video capture card used has acquisition parameters of 1080 / 60Hz and a data transmission interface of USB 3.0, backward compatible with USB 2.0.

[0059] Specifically, the physical components corresponding to the video stream acquisition module are a signal display and a video capture card.

[0060] Specifically, the physical components corresponding to the image capture module, the noise reduction and binarization module, the key point monitoring module, and the video stream annotation module are computing servers. The computing server processes each frame of image collected by the video capture card and uses algorithms to detect whether key nodes appear.

[0061] Specifically, the physical counterpart to the video stream display module is a monitoring results display. This display shows the results processed by the computing server on the screen, making it convenient for surgeons to observe.

[0062] Example 2

[0063] Based on Example 1, the key point monitoring module pre-sets the key points as the start point and peak point of each beat signal, and calculates the peak time of each beat signal based on the start point and peak point. In real-world scenarios, the detection of multiple signals may be involved, requiring well-defined rules for pixel-by-pixel detection. The pre-set threshold range in the video stream annotation module is a set range for the peak time. If the peak time exceeds the set range, the corresponding position needs to be marked with a bright color in the video stream, and an audio prompt should be given; otherwise, the original image should be returned.

[0064] Example 3

[0065] like Figure 3 As shown, a non-invasive ECG signal-assisted monitoring and alarm method provided by an embodiment of the present invention includes the following steps:

[0066] Step 1: Use a video capture card to acquire a real-time video stream of the entire screen content from the electrophysiological recording system;

[0067] Step 2: Extract the signal image from the pre-defined position in the acquired video stream;

[0068] Step 3: Preprocess the captured signal image, including removing noise points and binarizing the image, to obtain a clean signal image;

[0069] Step 4: For the preprocessed signal image, use the peak time detection algorithm to detect key points and determine if there are key points; if there are key points, mark the location of the key points and provide sound prompts.

[0070] Step 5: Display the signal image as a video stream on the results display screen for real-time observation by the user.

[0071] Example 4

[0072] Building upon Example 3, step 2 requires knowing the position of each signal within each frame of the image. The pre-defined positions in step 2 include y_offset, x_offset, height, and width. Here, y_offset refers to the pixel's offset in height (from top to bottom), x_offset refers to the pixel's offset in width (from left to right), and y_offset and x_offset determine the starting point. height represents the height of the image to be cropped, and width represents the width of the image to be cropped. Height and width determine the range of the cropped image. The purpose of cropping the signal area is to eliminate interference from irrelevant pixels, such as non-signal areas at the interface edges and other signal areas. The more accurate the cropped image, the less time is required for subsequent analysis, and the more accurate the detection of key nodes.

[0073] Specifically, the signal image captured in step 2 is in RGB format.

[0074] Specifically, during the image capture process in step 2, in order to capture all pixels of the signal to be analyzed, pixels from other signals may be included. Some of these signal pixels are different colors, and some are not connected to the pixels of the current signal; these pixels are noise points. Step 3, noise removal, involves pre-setting thresholds on the three different color channels of the signal image to remove pixels of different colors, thereby filtering out noise points. Then, points not in connected regions are removed, and a connected component algorithm is used to extract the signal to be analyzed.

[0075] Specifically, the binarized image in step 3 will also calculate the position pos of the first white pixel in each column and the distance value from the first white pixel to the bottom of the signal image.

[0076] Example 5

[0077] Based on Example 4, step 4 includes the following sub-steps:

[0078] Step 4.1: Based on the maximum value of the pixel detection area, obtain the peak position, including calculating the average value of the distance value, denoted as average, and use the find_peaks function in the scipy.signal module to find the local maximum value of the signal in the signal image;

[0079] Step 4.2: Remove redundant peaks, including removing peaks in the first 30 pixels and the last 12 pixels; remove small peaks, including removing peaks with an average drop of less than 3.5, where the drop refers to the average difference between the peak value and the surrounding 12 pixels; remove isolated peaks, including traversing the peak list and adding peaks with a distance between adjacent peaks between (5, 100) to the peaks_add_list list; return the peaks_add_list list after deduplication and sorting, which contains the peak indices that meet the criteria.

[0080] Step 4.3: Detection of biphasic time difference. Large and small peaks appear in pairs according to a certain frequency during cardiac pacing. In electrophysiological recording systems, leads V1-V6 typically exhibit a regular biphasic pattern. After calculating the peak positions in Step 4.1, the positions corresponding to each pair of peaks are obtained, denoted as pos_pair = [x10, x11, x20, x21, x30, x31...], where x10 and x11 are the positions of a pair of large and small peaks, x20 and x21 are the positions of a pair of large and small peaks, and so on. Subtracting each pair yields the time interval between the two peaks, denoted as t = [t1, t2, t3...].

[0081] Step 4.4: Mark the double-peak time difference detection results, including marking the double-peak time difference on the signal image, marking nodes that exceed the threshold with a conspicuous color, and giving a warning including an audible prompt.

[0082] Step 4.5: The His bundle (H) lead in the heart often exhibits multiple peaks clustered together, appearing at a certain frequency according to cardiac pacing. Multi-peak separation detection includes grouping the peaks into clusters and selecting the dominant peak within each cluster, recording it in the Not_Split library, and then determining whether split peaks exist, recording them in the Split library.

[0083] Step 4.6: Mark the multi-peak separation detection results, including marking all detected peaks on the image. If a split peak is detected, mark it with a conspicuous color and provide an audio prompt.

[0084] Specifically, in step 4.5, the standard for grouping peaks is to group those within 100 pixels into one group; the standard for determining whether a peak is split is that the distance between two peaks is between 20 and 50 pixels, the numerical difference is less than 10 pixels, and the value of one of the peaks is greater than 140 pixels.

[0085] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A non-invasive ECG signal-assisted monitoring and alarm method, characterized in that, The non-invasive ECG signal-assisted monitoring and alarm method includes the following steps: Step 1: Use a video capture card to acquire a real-time video stream of the entire screen content from the electrophysiological recording system; Step 2: Extract the signal image from the pre-defined position in the acquired video stream; Step 3: Preprocess the captured signal image, including removing noise points and binarizing the image to obtain a clean signal image; the binarized image will also calculate the position pos of the first white pixel in each column and the distance value from the first white pixel to the bottom of the signal image; Step 4: For the preprocessed signal image, use the peak time detection algorithm to detect key points and determine if there are key points; if there are key points, mark the location of the key points and provide sound prompts. Step 4 includes the following sub-steps: Step 4.1: Based on the maximum value of the pixel detection area, obtain the peak position, including calculating the average value of the distance value, denoted as average, and using the find_peaks function in the scipy.signal module to find the local maximum value of the signal in the signal image; Step 4.2: Remove redundant peaks, including removing peaks in the first 30 pixels and the last 12 pixels; remove small peaks, including removing peaks with an average left-right drop of less than 3.5, where the drop refers to the average difference between the peak value and the surrounding 12 pixels; remove isolated peaks, including traversing the peak list and adding peaks with a distance between adjacent peaks between (5, 100) to the peaks_add_list; return the peaks_add_list after deduplication and sorting, which contains the peak indices that meet the criteria. Step 4.3, Double Peak Time Difference Detection: After calculating the peak positions in Step 4.1, the positions corresponding to each pair of peaks are obtained, denoted as pos_pair = [x10, x11, x20, x21, x30, x31...], where x10 and x11 are a pair of large and small peak positions, x20 and x21 are a pair of large and small peak positions, and so on; subtracting each pair yields the time interval between the two peaks, denoted as t = [t1, t2, t3...]. Step 4.4: Mark the double-peak time difference detection results, including marking the double-peak time difference on the signal image, marking nodes that exceed the threshold with a conspicuous color, and providing an audio prompt; Step 4.5, Multi-peak separation detection, includes grouping peaks into clusters and selecting the main peak within each cluster, recording it in the Not_Split library, and then determining whether there are split peaks, recording them in the Split library; the standard for grouping peaks into clusters is that peaks within 100 pixels are considered as one cluster; the standard for determining whether a peak is split is that the distance between two peaks is between 20 and 50 pixels, the numerical difference is less than 10 pixels, and the value of one of the peaks is greater than 140 pixels; Step 4.6: Mark the multi-peak separation detection results, including marking all detected peaks on the image. If split peaks are detected, mark them with a conspicuous color and provide an audio prompt. Step 5: Display the signal image as a video stream on the result display screen for real-time observation by the user.

2. The non-invasive ECG signal-assisted monitoring and alarm method as described in claim 1, characterized in that, The preset position in step 2 includes y_offset, x_offset, height, and width. Here, y_offset refers to the offset of a pixel in height, x_offset refers to the offset of a pixel in width, y_offset and x_offset determine the starting point, height represents the height of the image to be cropped, and width represents the width of the image to be cropped. Height and width determine the range of the image to be cropped.

3. The non-invasive ECG signal-assisted monitoring and alarm method as described in claim 1, characterized in that, The signal image captured in step 2 is in RGB format. The noise removal in step 3 includes setting thresholds in advance on the three different color channels of the signal image, removing pixels of different colors to filter out noise points, removing points not in the connected region, and using a connected component algorithm to extract the signal to be analyzed.

4. A non-invasive ECG signal-assisted monitoring and alarm system, characterized in that, The non-invasive ECG signal-assisted monitoring and alarm method as described in any one of claims 1-3 includes a video stream acquisition module, an image capture module, a noise reduction and binarization module, a key point monitoring module, a labeled video stream module, and a video stream display module. The video stream acquisition module acquires the video stream from the electrophysiological recording system; The image cropping module is connected to the video stream acquisition module and extracts the signal image at a preset position from the acquired video stream; The denoising and binarization module is connected to the image cropping module and preprocesses the cropped signal image, including removing noise points and binarizing the image, to obtain a clean signal image. The key point monitoring module is connected to the denoising and binarization module, and detects the detection signals of pre-set key points by pixel in the pre-processed signal image. The labeled video stream module is connected to the key point monitoring module. If the detection signal of the key point exceeds the preset threshold range, the corresponding position is marked with a bright color in the signal image and a prompt is given. The video stream display module is connected to the labeled video stream module and displays the signal image on the result display screen in the form of a video stream, which is convenient for users to observe in real time.

5. The non-invasive ECG signal-assisted monitoring and alarm system as described in claim 4, characterized in that, The video stream acquisition module acquires the video stream from the electrophysiological recording system through a video capture card, connects it to the computing server via a USB interface, and acquires the video stream frame by frame using the VideoCapture class of the OpenCV library.

6. The non-invasive ECG signal-assisted monitoring and alarm system as described in claim 5, characterized in that, The video capture card is used to transmit the real-time images of the electrophysiological recording system frame by frame to the computing server. The video capture card has acquisition parameters of 1080 / 60Hz and a data transmission interface of USB 3.0, which is backward compatible with USB 2.

0.

7. The non-invasive ECG signal-assisted monitoring and alarm system as described in claim 4, characterized in that, The key point monitoring module pre-sets the key points as the starting point and peak point of each beat signal, and calculates the peak time of each beat signal based on the starting point and the peak point.

8. The non-invasive ECG signal-assisted monitoring and alarm system as described in claim 7, characterized in that, The threshold range preset in the labeled video stream module is a set range regarding the peak time.

Images