Online detection method and quantitative analysis system of target signal

Abstract

The present invention belongs to the field of physiological electrical signal detection technology, and specifically relates to an online detection method and quantitative analysis system for a target signal. The online detection method for a target signal includes: setting a characteristic parameter threshold of a template waveform; performing waveform detection in a global online signal based on the characteristic parameter threshold to obtain a first waveform signal; wherein the characteristic parameters of the template waveform include: peak amplitude, waveform time limit, amplitude difference between the waveform start point and the maximum peak point, amplitude difference between the waveform end point and the maximum peak point, slope from the waveform start point to the maximum peak point, and slope from the waveform end point to the maximum peak point. By setting the characteristic parameter threshold of the template waveform in the waveform detection model, the model performance is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of physiological electrical signal detection, and in particular relates to an online detection method and a quantitative analysis system of a target signal. Background Art

[0002] EEG signals can reflect normal and abnormal physiological activities of the brain. By identifying and analyzing specific waveforms, they can be used to understand the functional state of the brain, diagnose neurological diseases, monitor treatment effects, etc. They are widely used in the fields of neurology and clinical neurophysiology. EEG signals usually collect a large number of leads, take a long time to collect, and require a lot of work when manually identifying them. Therefore, waveform detection algorithms based on EEG signals are increasingly attracting people's attention. At present, there are three main deficiencies in the detection of EEG signals: First, in clinical use, EEG signals are mainly identified and judged manually, which requires professional physicians to have rich clinical experience, is time-consuming and labor-intensive, and lacks objectivity; second, the current automatic detection algorithm requires manual cooperation in the early training data to mark waveforms, which requires a lot of work in the early stage; third, there is a lack of intuitive quantitative analysis of the waveform detection results of EEG signals. Therefore, software recognition algorithms are used in related technologies to detect target waveforms. For example, patent number CN201910511248.7 discloses an electrophysiological testing device for an automated auditory brainstem implant. It is mentioned in

[0042] -

[0045] of the specification that in step S23, the eABR waveform is automatically identified by the software recognition algorithm module in the second computer 2. Among them, the starting point of the eABR waveform generally appears within 1ms and the entire eABR waveform time limit is approximately within 3ms, so the waveform can be automatically identified by the software recognition algorithm module within the eABR waveform time limit. The software recognition algorithm module also differentiates the slope of the data points of the eABR waveform to find the starting point, peak, and trough of the waveform, and then locates and identifies the entire eABR waveform, and automatically calculates the latency, amplitude, time limit and other data of the eABR waveform. Using the detected electrical stimulation auditory brainstem evoked potential to judge the electrode response can effectively improve the efficiency of the audiologist's electrode testing during surgery, so the accuracy of waveform recognition is the main consideration of this technical solution. However, using the waveform detection results as training data for machine models requires taking into account both the comprehensiveness and accuracy of the data. As we all know, in the process of software identifying the target waveform, an excessively large waveform parameter threshold leads to an increased false positive rate, while an excessively small waveform parameter threshold leads to an increased missed positive rate. Therefore, how to set the waveform parameter threshold to balance the false positive rate and missed positive rate and improve the performance of the detection model is still an urgent problem to be solved. Summary of the invention

[0003] The present invention provides an online detection method and a quantitative analysis system for a target signal, so as to improve the model performance by setting a characteristic parameter threshold of a template waveform.

[0004] In order to solve the above technical problems, the present invention provides an online detection method for a target signal, comprising: setting a characteristic parameter threshold of a template waveform; performing waveform detection in a global online signal based on the characteristic parameter threshold to obtain a first waveform signal; wherein the characteristic parameters of the template waveform include: the peak amplitude of the template waveform, the waveform time limit, the amplitude difference between the waveform starting point and the maximum peak point, the amplitude difference between the waveform ending point and the maximum peak point, the slope from the waveform starting point to the maximum peak point, and the slope from the waveform ending point to the maximum peak point.

[0005] Furthermore, the waveform start point and waveform end point are obtained by a peak detection method, that is, the waveform start point and waveform end point of the target signal are found respectively according to the change in waveform direction of the target signal on both sides of the waveform peak point.

[0006] Further, a single waveform signal is set as the template waveform of the target signal; multiple waveform signals including the template waveform are extracted from the local online signal; the characteristic parameter values ​​of all waveform signals are obtained and sorted by size; the percentage α of each characteristic parameter of the template waveform in the sorting position in the corresponding category is obtained i , (1-α i The product of ) and the corresponding feature parameter is taken as the corresponding feature parameter threshold.

[0007] Furthermore, it also includes: waveform screening, screening waveform signals based on the zero-crossing rate threshold after signal data differentiation; waveform extraction, online extraction of waveform signals to obtain a target signal set; template matching, screening waveform signals that match the template waveform from the target signal set based on a waveform similarity threshold.

[0008] Furthermore, when the target signal is configured as a composite waveform signal, that is, composed of a template waveform and a template residual waveform, the target signal detection method also includes: residual waveform comparison; the residual waveform comparison includes: setting a characteristic parameter threshold of the template residual waveform; taking the position of the template waveform as a reference, extracting a second waveform signal with the same waveform time limit as the target signal and not containing the first waveform signal; comparing the second waveform signal with the template residual waveform to filter the first waveform signal according to the characteristic parameter threshold of the template residual waveform; wherein the characteristic parameters of the template residual waveform include: peak amplitude and waveform time limit.

[0009] Furthermore, it also includes: constructing a waveform detection model; constructing a database, constructing a database of the target signal waveform using the first waveform signal; model training, using the database to train the waveform detection model; updating the database, updating the database with manual labeling results to optimize the waveform detection model.

[0010] In a second aspect, the present invention provides a quantitative analysis system for a target signal, comprising: a processor, which runs the online detection method to detect the target signal; a quantitative analysis module, comprising at least one of a discharge frequency quantitative analysis module and a multi-aspect indicator statistical analysis module; and a display module, which is used to display the analysis results of the quantitative analysis module on the target signal.

[0011] Furthermore, the discharge frequency quantitative analysis module is suitable for mapping the discharge frequency of each lead to a 3D brain topography map to form an analysis result.

[0012] Furthermore, the multi-directional indicator statistical analysis module is suitable for measuring the analysis results through at least one quantitative analysis indicator: the quantitative analysis indicators include: the ratio of the number of target signals to background signals, the time limit ratio of the target signal to the background signal, the frequency of the target signal, the number of target signal detections per unit time, the proportion of the target signal to the overall time length, the average energy of the target signal, and at least one of the spatial synchronization coefficients of the target signal.

[0013] In a third aspect, the present invention provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the method when executed by a processor.

[0015] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which implements the steps of the method when executed by a processor.

[0016] The beneficial effect of the present invention is that the online detection method and quantitative analysis system of the target signal of the present invention can accurately lock the target waveform, balance the accuracy and missed detection rate, and improve the model performance by setting the characteristic parameter threshold of the template waveform in the waveform detection model, and using six characteristic parameters such as peak amplitude, waveform time limit, amplitude difference between the waveform starting point and the maximum peak point, amplitude difference between the waveform ending point and the maximum peak point, slope from the waveform starting point to the maximum peak point, and slope from the waveform ending point to the maximum peak point to perform waveform detection in the global online signal.

[0017] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 It is a detection flow chart of a single waveform signal.

[0021] Figure 2 This is a schematic diagram of the operation interface for selecting a template waveform in a single waveform signal.

[0022] Figure 3 It is a schematic diagram of the operation interface for detecting results in a single waveform signal.

[0023] Figure 4 It is a detection flow chart of composite waveform signal.

[0024] Figure 5 It is a schematic diagram of the operation interface for detecting results from the combined waveform signal.

[0025] Figure 6 It is a flow chart of fully automatic online detection of a single waveform signal.

[0026] Figure 7 It is a statistical histogram of the discharge frequency of the detection results in each lead.

[0027] Figure 8 It is a statistical topographic map of the discharge frequency detected in each lead.

[0028] Fig. 9 This is a graph showing the quantitative analysis results of the detection results in lead C3.

[0029] Fig.10 This is a graph showing the quantitative analysis results of the detection results in lead C4.

[0030] Fig.11 It is a schematic diagram of the operation interface for selecting the template waveform in the C4 lead in the test example.

[0031] Fig.12 This is a diagram showing the local detection results of the target signal under test condition 1.

[0032] Fig.13 This is a diagram showing the local detection results of the target signal under test condition 2.

[0033] Fig.14 This is a diagram showing the local detection results of the target signal under test condition 3. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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.

[0035] Example 1

[0036] See Figure 1 , this embodiment 1 provides an online detection method for a target signal, comprising: setting a characteristic parameter threshold of a template waveform; performing waveform detection in a global online signal based on the characteristic parameter threshold to obtain a first waveform signal; wherein the characteristic parameters of the template waveform include: peak amplitude of the target signal, waveform time limit, amplitude difference between the waveform start point and the maximum peak point, amplitude difference between the waveform end point and the maximum peak point, slope from the waveform start point to the maximum peak point, and slope from the waveform end point to the maximum peak point.

[0037] In this case, the waveform to be detected can be locked by the six characteristic parameters of the template waveform to balance the relationship between accuracy and missed detection rate. Specifically, peak amplitude: limit the amplitude of the waveform, and perform peak detection of the waveform based on the peak amplitude to obtain the initial waveform and its corresponding start and end points; waveform time limit: limit the width of the target waveform; amplitude difference: because different types of waveforms have different relative lengths from the start point and end point to the peak point, by comparing the amplitude difference from the start point and end point to the peak point, some non-target waveforms can be removed, such as the distinction between spike waves that often appear in epileptic EEG signals and normal physiological waveforms superimposed on slow waves, and some normal physiological waveforms superimposed on slow waves. It will also make the amplitude and time limit close to the spike wave, but the start and end points are on the slow wave. The two can be distinguished by calculating the amplitude difference from the start and end points to the peak point; Slope: The rise and fall speeds of different waveforms are different. By limiting the slope, some waveforms that are too gentle or too sharp can be effectively removed. For example, the distinction between spike waves and sharp waves that often appear in epileptic EEG signals. The two are close in amplitude and amplitude difference, which can easily cause false detection, but sharp waves are smoother than spike waves, and the rise and fall of the waveform are slower. Using slope to limit can effectively reduce the occurrence of false detection.

[0038] Optionally, the waveform start point and waveform end point are obtained by a peak detection method, that is, according to the change in the waveform direction of the target signal on both sides of the waveform peak point, the waveform start point and waveform end point of the target signal are found respectively. Specifically, all peak points that meet the requirements are detected first, and the differential data of the data is calculated. Starting from the location of the peak point, the location of the data point where the numerical sign changes is searched forward in the differential data, that is, the point where the data waveform direction changes, and the data continuity threshold is set. If the direction of the data continuity threshold points does not change from the direction change point forward, then the point where the current data direction changes is found as the start point, and the end point is obtained in the same way.

[0039] Optionally, the template waveform is configured as a single waveform signal, which can be manually selected. Therefore, the online detection method of the target signal in this embodiment 1 can be considered as a semi-automatic detection method supervised by humans, and the target signal can be detected online by using the semi-automatic waveform detection model of the software.

[0040] Furthermore, the online detection method also includes: waveform screening, screening the waveform signal based on the zero-crossing rate threshold after signal data differentiation, that is, calculating the zero-crossing rate of the differential data of the first waveform signal, comparing it with the zero-crossing rate threshold of the template waveform, and eliminating the artifact waveform greater than the zero-crossing rate threshold. Waveform extraction, online extraction of waveform signals to obtain a target signal set. Template matching, based on the waveform similarity threshold, filtering the waveform signal that matches the template waveform from the target signal set, saving the waveform greater than the waveform similarity threshold, and eliminating the waveform containing noise.

[0041] In this case, the database is constructed by constructing the first waveform signal or the target signal set after waveform screening and template matching, which can be used as training data for the waveform detection model to replace manual labeling and reduce the initial workload. In addition, the labels can be refreshed and the database can be updated based on the evaluation results after automatic detection to further train and optimize the deep model, realize the closed loop of data and model, and obtain more objective and accurate recognition results.

[0042] In this case, the target signal to be detected can be a single waveform signal or a composite waveform signal. The specific detection process is as follows:

[0043] Optionally, when the target signal is configured as a single waveform signal, the target signal is equivalent to the template waveform. The template waveform requires the user to mark the template waveform online or select the template waveform from a database, extract characteristic thresholds from the four aspects of the template waveform: amplitude, time limit, amplitude difference and slope, extract the waveform, and screen the waveform according to the template matching algorithm to achieve detection of the waveform in the target signal. Taking the online marking of the template waveform as an example, the specific process of setting the characteristic parameter threshold of the template waveform includes: the user selects the target signal template waveform to be detected, and each target signal lead can independently select one or more; extract multiple waveform signals including the template waveform from any lead in the local online signal (that is, the channel to which the acquisition signal belongs), such as Figure 2 As shown in the highlighted waveform, the corresponding single waveform signals are selected as template waveforms in lead C3, lead C4, and lead P3 respectively; the characteristic parameter values ​​of all waveform signals are obtained and sorted by size; the percentage α of each characteristic parameter in the sorting position is obtained i , and (1-α i ) and the corresponding characteristic parameter as the characteristic parameter threshold; based on the characteristic parameter threshold, waveform detection is performed in the global online signal to obtain a first waveform signal, such as Figure 3 Specifically, the percentage of each characteristic parameter in the sorting position α i It is the percentage of the corresponding characteristic parameter of the template waveform in the local online signal. For example, if the peak amplitude is 125uV, sort the peak values ​​of all waveforms in the local online signal from small to large. If the peak value of the template waveform is at 20%, then the percentage of the peak amplitude α i That is 20%, and the peak amplitude threshold of the target signal is 80%*125uV=100uV. Similarly, other characteristic parameter thresholds are set in the same way. In the process of setting the characteristic parameter threshold of the template waveform, you can also first verify the detection effect of the characteristic parameter threshold in the local online signal, and manually adjust and optimize the characteristic parameter threshold to detect all target signals in the local online signal, and then retrieve the target signal in the global online signal.

[0044] Optional, see Figure 4, when the target signal is configured as a composite waveform signal, that is, it is composed of a template waveform and a template residual waveform, the target signal detection method, based on the first detection method, also includes: residual waveform comparison; the residual waveform comparison includes: setting the characteristic parameter threshold of the template residual waveform; taking the position of the template waveform as a reference, extracting a second waveform signal with the same waveform time limit as the target signal and not containing the first waveform signal; comparing the second waveform signal with the template residual waveform to filter the first waveform signal according to the characteristic parameter threshold of the template residual waveform; wherein the characteristic parameters of the template residual waveform include: peak amplitude, waveform time limit. In this case, the characteristic parameter threshold of the template residual waveform is similar to the characteristic parameter threshold setting method of the template waveform, and will not be repeated.

[0045] Obviously, the target signal detection method in this case uses a single waveform signal as a template waveform, extracts multiple waveform signals including the template waveform, and obtains the characteristic parameter threshold of the template waveform by the percentage of the characteristic parameter in the sorting position in the corresponding classification, which can realize the waveform detection in the target signal, reduce the time spent, and ensure the accuracy of the detection result; at the same time, the matching and screening of the template residual waveform takes into account the detection needs of the composite waveform signal. Figure 5 As shown in the figure, the target signal of lead F3 is configured as a composite waveform, the target signal of lead C4 is configured as a mixture of a single waveform and a composite waveform, and the target signals of lead T3, lead P4, etc. are configured as a single waveform. Therefore, the requirements of each lead for different waveforms of the target signal can be fully met.

[0046] Example 2

[0047] See Figure 6 On the basis of Example 1, this Example 2 further provides an online detection method for a target signal, further comprising: constructing a waveform detection model; constructing a database, constructing a database of the target signal waveform with a first waveform signal; model training, using the database to train the waveform detection model; updating the database, updating the database with the manual label result to optimize the characteristic parameter threshold of the template waveform and improve the performance of the waveform detection model. In the detection process of the target signal, the waveform detection model and its corresponding database are constructed with the help of the detection result of Example 1, that is, the target signal set, and the data label is refreshed by expert marking to realize the training of the waveform detection model, so as to realize the fully automatic online detection of the target signal.

[0048] Example 3

[0049] Based on Example 1 or 2, this Example 3 provides a quantitative analysis system for a target signal, including: a processor, running the online detection method to obtain the target signal; a quantitative analysis module, including at least one of a discharge frequency quantitative analysis module and a multi-dimensional indicator statistical analysis module; a display module, used to display the analysis results of the quantitative analysis module on the target signal.

[0050] Optionally, the discharge frequency quantitative analysis module is suitable for counting the number of discharges of each lead and drawing a histogram of frequency statistics (such as Figure 7 ), and mapped the results onto a 3D brain topography map (as Figure 8 shown).

[0051] Further, the multi-directional index statistical analysis module is suitable for measuring the analysis results through at least one quantitative analysis index: the quantitative analysis index includes: the ratio of the number of target signals to background signals, the time limit ratio of the target signal to the background signal, the frequency of the target signal, the number of target signal detections per unit time, the proportion of the target signal to the overall time, the average energy of the target signal, and at least one of the spatial synchronization coefficients of the target signal. The meanings of the quantitative analysis indicators are shown in Table 1 below.

[0052] Table 1 Meaning of quantitative analysis indicators

[0053]

[0054] When displaying the actual results, the seven quantitative analysis indicators can be used to draw a bar graph of each indicator, and the quantitative analysis indicator results for each hour can be counted. The template waveform, quantitative analysis indicators, and 3D brain topography of the lead detection results can be displayed on the same operation interface to obtain more intuitive and accurate analysis results, which is convenient for users to use the software for operation and manual supervision. It can realize the detection and quantitative analysis of single wave manual supervision, and the detection and quantitative analysis of complex wave manual supervision. For example, the full detection results of the target waveform in the C3 lead are as follows: Fig. 9 As shown in the figure, the complete detection results of the target waveform in lead C4 are as follows Fig.10 shown.

[0055] Example 4

[0056] Based on Embodiment 1 or 2, Embodiment 4 provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0057] Example 5

[0058] On the basis of Embodiment 1 or 2, this Embodiment 5 provides a computer-readable storage medium on which a computer program is stored, and the computer program implements the steps of the method when executed by a processor.

[0059] Example 6

[0060] On the basis of Embodiment 1 or 2, this Embodiment 6 provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method are implemented.

[0061] Test example

[0062] The single waveform signal is set as the template waveform of the target signal. Take the template waveform of lead C4 as an example. Fig.11 The waveform is shown in the box.

[0063] Test condition 1: The characteristic parameters of the template waveform are set to the peak amplitude and waveform time limit of the template waveform; (1-α i The product of ) and the corresponding characteristic parameter is used as the corresponding characteristic parameter threshold, and the online detection method of this embodiment 1 is used to search in the local online data to verify the detection result. The local detection result is as follows Fig.12 This is shown in the highlighted waveform.

[0064] Test condition 2: The difference from test condition 1 is that the characteristic parameters of the template waveform are the peak amplitude of the template waveform, the waveform time limit, the amplitude difference between the waveform start point and the maximum peak point, the amplitude difference between the waveform end point and the maximum peak point, the slope from the waveform start point to the maximum peak point, and the slope from the waveform end point to the maximum peak point; the other conditions are the same to verify the detection results. The local detection results are as follows Fig.13 This is shown in the highlighted waveform.

[0065] Test condition 3: The difference from test condition 2 is that the characteristic parameter value of the template waveform is used as the corresponding characteristic parameter threshold. The local detection results are as follows: Fig.14 This is shown in the highlighted waveform.

[0066] Test results: As shown in Table 2, the following conclusions can be drawn:

[0067] (1) Comparing the detection results of test conditions 1 and 2, it can be seen that the missing amplitude difference and slope, i.e., the amplitude difference from the waveform start point to the maximum peak point, the amplitude difference from the waveform end point to the maximum peak point, the slope from the waveform start point to the maximum peak point, and the slope from the waveform end point to the maximum peak point, as the characteristic parameters of the template waveform, will lead to overly broad waveform detection conditions, increase the false detection of the target waveform, and thus maximize the local detection. Therefore, the global search is performed using this test condition as the characteristic parameter of the template waveform, and the result is the same. At the same time, in the treatment of EEG epileptic seizures, if the global search is performed using test condition 1, the target signal set is obtained and used for machine training, it will inevitably lead to a significant increase in the false detection rate of the signal detection model, and a large number of cases will occur where electrical stimulation is performed to suppress epilepsy during non-epileptic seizures.

[0068] (2) By comparing the detection results of test conditions 2 and 3, it can be seen that using the characteristic parameter value of the template waveform as the corresponding characteristic parameter threshold will result in the local detection result only detecting waveforms that are exactly the same as the template waveform. Although the target signal can be accurately detected, the signal that is close to the template waveform will be missed, resulting in an increase in missed detections and a decrease in local detections. When judging EEG epileptic seizures, these signals that are close to the template waveform may also be EEG epileptic seizure signals. If a global search is performed under test condition 3 to obtain the target signal set and use it for machine training, the missed detection rate of the signal detection model will inevitably increase significantly, which is not conducive to the judgment of epileptic seizures. Ultimately, during the treatment of EEG epileptic seizures, there will be a large number of cases where electrical stimulation is not implemented to suppress epilepsy during epileptic seizures.

[0069] Table 2 Comparison of detection results of single waveform signal

[0070] Test conditions Local detection Missed detection False positives Test condition 1 34 0 10 Test condition 2 24 0 0 Test condition 3 4 20 0

[0071] In the description of the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0072] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0073] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0074] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An online detection method for a target signal, characterized in that: include: Set the characteristic parameter threshold of the template waveform; Performing waveform detection in the global online signal based on the characteristic parameter threshold to obtain a first waveform signal; in The characteristic parameters of the template waveform include: peak amplitude of the template waveform, waveform time limit, amplitude difference between the waveform start point and the maximum peak point, amplitude difference between the waveform end point and the maximum peak point, slope from the waveform start point to the maximum peak point, and slope from the waveform end point to the maximum peak point.

2. The online detection method according to claim 1, characterized in that: The waveform start point and waveform end point are obtained by a peak detection method, that is, the waveform start point and waveform end point of the target signal are found respectively according to the change in the waveform direction of the target signal on both sides of the waveform peak point.

3. The online detection method according to claim 1, characterized in that: The step of setting the characteristic parameter threshold of the template waveform comprises: Set a single waveform signal as the template waveform of the target signal; Extracting multiple waveform signals including a template waveform from local online signals; Get the characteristic parameter values ​​of all waveform signals and sort them by size; Get the percentage α of each characteristic parameter of the template waveform in the corresponding classification i , (1-α i The product of ) and the corresponding feature parameter is taken as the corresponding feature parameter threshold.

4. The online detection method according to claim 1, characterized in that: Also includes: Waveform screening: Screening waveform signals based on the zero-crossing rate threshold after signal data differentiation; Waveform extraction: extract waveform signals online to obtain the target signal set; Template matching, based on the waveform similarity threshold, filters the waveform signals that match the template waveform from the target signal set.

5. The online detection method according to claim 1, characterized in that: When the target signal is configured as a composite waveform signal, namely, composed of a template waveform and a template residual waveform, the target signal detection method further includes: residual waveform comparison; The residual waveform comparison includes: Set the characteristic parameter threshold of the template residual waveform; Taking the position of the template waveform as a reference, extracting a second waveform signal having the same waveform time limit as the target signal and not containing the first waveform signal; The second waveform signal is compared with the template residual waveform to filter the first waveform signal according to the characteristic parameter threshold of the template residual waveform; wherein The characteristic parameters of the template residual waveform include: peak amplitude and waveform time limit.

6. The online detection method according to claim 1, characterized in that: Also includes: Build a waveform detection model; Building a database, using the first waveform signal to build a database of the target signal waveform; Model training, using the database to train the waveform detection model; Update the database with manual labeling results to optimize the waveform detection model.

7. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A quantitative analysis system for a target signal, characterized in that: include: A processor, executing the online detection method according to any one of claims 1 to 6 to detect a target signal; A quantitative analysis module, including at least one of a discharge frequency quantitative analysis module and a multi-directional index statistical analysis module; The display module is used to display the analysis results of the quantitative analysis module on the target signal.

11. The quantitative analysis system according to claim 10, characterized in that: The discharge frequency quantitative analysis module is suitable for mapping the discharge frequency of each lead to a 3D brain topography map to form an analysis result.

12. The quantitative analysis system according to claim 10, characterized in that: The multi-faceted indicator statistical analysis module is suitable for measuring the analysis results through at least one quantitative analysis indicator: The quantitative analysis indicators include: the ratio of the number of target signals to background signals, the time limit ratio of the target signal to the background signal, the frequency of the target signal, the number of target signal detections per unit time, the proportion of the target signal to the overall time length, the average energy of the target signal, and at least one of the spatial synchronization coefficients of the target signal.

Images