Online detection method of target signal, epilepsy detection system

Abstract

The present invention belongs to the field of signal detection technology, and specifically relates to an online detection method for a target signal and an epilepsy detection system. The online detection method for the target signal includes: signal detection to obtain a target signal set and its corresponding signal template; signal extraction, i.e., extracting the target signal using the waveform characteristics of the online signal; iterative optimization, i.e., adjusting the matching threshold between the target signal and the signal template according to the extraction result of the target signal to screen the target signal and update the target signal set. The detected characteristic parameters are corrected using the extraction result of the target signal to iteratively optimize the target signal set obtained by the signal detection. Iterative optimization can be performed online without the need for historical data or machine learning, and the target waveform can be extracted more comprehensively and accurately.

Description

Technical Field

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

[0002] Epilepsy is a common neurological disease characterized by recurrent seizures caused by excessive and abnormal neuronal activity in the cerebral cortex. EEG is a technique for measuring and recording brain activity and is widely used in the detection and analysis of epileptic seizures. However, it is often difficult and inefficient to identify subtle but critical changes in EEG waveforms through visual inspection, so the development of automatic detection and quantitative analysis algorithms for epilepsy is particularly important.

[0003] The detection of abnormal discharges of epilepsy based on EEG signals is the most commonly used method. In the related art, EEG can be identified by automatic detection algorithms, but due to the limitations of historical data, only the discharge waveforms corresponding to one or several historical templates can be detected. At the same time, historical data and online data are also essentially different, resulting in the detection results of epilepsy being not comprehensive and accurate enough. For example, patent CN 110811609 A discloses an intelligent detection device for epilepsy spikes based on the fusion of adaptive template matching and machine learning algorithms, which fuses the spike detection results obtained by the adaptive template matching spike detection module and the spike detection results obtained by the machine learning spike detection module. If the adaptive template matching spike detection module and the machine learning spike detection module detect spike signals in the same segment, the final result is recorded as the presence of spike signals in the segment, and it is regarded as an epileptic spike. Although the accuracy of the detection results can be improved by combining the selected template with historical data for verification, the limitations and differences of historical data have not been solved. Summary of the invention

[0004] The invention provides an online detection method of a target signal and an epilepsy detection system to improve the comprehensiveness of the detection waveform.

[0005] In order to solve the above technical problems, the present invention provides an online detection method for a target signal, comprising: signal detection to obtain a target signal set and its corresponding signal template; signal extraction, i.e. extracting the target signal using the waveform characteristics of the online signal; iterative optimization, i.e. adjusting the matching threshold between the target signal and the signal template according to the extraction result of the target signal to screen the target signal and update the target signal set.

[0006] Furthermore, the signal detection includes any one of a first detection method, a second detection method, and a third detection method; wherein the first detection method is configured to obtain a first signal template by online detecting the target signal based on preset characteristic parameters; the second detection method is configured to detect the target signal online based on a second signal template selected based on the signal waveform; and the third detection method is configured to obtain the target signal by combining the first detection method and the second detection method.

[0007] Furthermore, the first detection method includes: initially setting characteristic parameters of the target signal; local pre-adjustment, i.e., adjusting the characteristic parameters in the local online signal to change the extraction result of the target signal; global retrieval, i.e., detecting the target signal according to the characteristic parameters in the global online signal to obtain a first target signal set; waveform clustering of the first target signal set to obtain a waveform template as the first signal template.

[0008] Furthermore, the characteristic parameters include at least one of a time period, an amplitude threshold, an energy threshold, an amplitude time limit, and an energy time limit.

[0009] Further, when the signal detection is configured as the first detection mode, the iterative optimization is configured as template matching, including: detecting the matching degree between the target signal in the target signal set and the signal template, removing the target signal with the lowest matching degree and below the matching degree threshold in the target signal set, updating the target signal set and updating the signal template with the average of the remaining target signals, repeating the foregoing operations until the matching degree of all target signals in the target signal set and the updated signal template is greater than the matching degree threshold; adjusting the matching degree threshold based on the extraction result of the target signal, and repeating the template matching until the target signal extraction result meets expectations.

[0010] Furthermore, the second detection method includes: obtaining a second signal template, that is, using the signal waveform diagram of the online signal to extract the target waveform as the second signal template; locally pre-adjusting the matching degree, that is, adjusting the matching degree threshold in the local online signal to change the extraction result of the target signal; global retrieval, that is, retrieving the target signal according to the matching degree threshold in the global online signal to obtain a second target signal set.

[0011] Further, when the signal detection is configured as the second detection mode, the iterative optimization is configured to adjust the matching threshold, including: adjusting the matching threshold based on the extraction result of the target signal, and repeating the template matching until all target signals in the target signal set meet the set extraction result of the target signal.

[0012] Furthermore, the third detection method includes: performing waveform merging on the target signal sets obtained by the first detection method and the second detection method to obtain a third target signal set; and performing waveform clustering on the third target signal set to obtain the third signal template.

[0013] Further, the signal extraction includes: obtaining the waveform position of the target signal, that is, selecting the start point and the end point of the target signal in the waveform display diagram; obtaining the channel to which the target signal belongs; and highlighting the target signal through the signal waveform diagram of the online signal.

[0014] In a second aspect, the present invention provides an epilepsy signal detection system, comprising: a processor, which runs the online detection method; a human-computer interaction machine connected to the processor, and whose operation interface is provided with function options, a signal extraction area, and quantitative analysis options; wherein a signal detection method option box is provided in the function option to select different methods for signal detection; the signal extraction area is suitable for displaying a signal waveform diagram of the online signal for extracting the target signal; the quantitative analysis option is suitable for performing quantitative analysis on the target signal in the target signal set to obtain a measurement indicator of the epilepsy signal.

[0015] Further, the quantitative analysis options include at least one of discharge frequency analysis, intensity analysis, and spatial time series analysis; wherein the discharge frequency analysis includes: selecting the required channel, statistically analyzing the discharge frequency of the target signal, calculating the frequency of global or certain type of discharge for each channel, and drawing a histogram and a 3D brain topography of the corresponding discharge frequencies of all channels; the intensity analysis includes: drawing a time-frequency-spatial energy spectrum for the target signal, selecting a display period, interval, frequency range, and display channel; drawing a 3D brain topography of the extracted waveform energy in each window and a time-frequency diagram of the display period; adjusting the time to display the waveform energy changes of the target signal at different times; the spatial time series analysis includes: taking the target signal set as a reference, intercepting the signal data of all channels in the same period, and the signal data of each period is a group; dividing the groups of waveforms with the same channel position corresponding to the maximum waveform into one category, calculating the mean of the waveforms of each channel in each category as a signal template for display; determining the peak moment corresponding to the maximum waveform of each type of signal template, and drawing a 3D brain topography with the voltage value of each channel at the current moment.

[0016] 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.

[0017] 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.

[0018] 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.

[0019] The beneficial effect of the present invention is that the online detection method of the target signal and the epilepsy detection system of the present invention extract the target signal with the help of the signal waveform diagram of the online signal, and adjust the matching threshold between the target signal and the signal template according to the extraction result of the target signal to screen the target signal and update the target signal set; the extraction result of the target signal is used to correct the required parameters of signal detection and signal extraction, and then the processor automatically performs signal detection and signal extraction on the online signal according to the required parameters, which can not only cover more target signals, but also perform online iterative optimization without the need for historical data or machine learning, avoiding the difference between historical data and online data, and can extract the target waveform more comprehensively and accurately.

[0020] 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 and the drawings.

[0021] 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

[0022] 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.

[0023] Figure 1 It is a workflow diagram of the online detection method.

[0024] Figure 2 It is a schematic diagram of the operation interface of the human-computer interaction machine of the epilepsy signal detection system.

[0025] Figure 3 This is the workflow diagram of the first detection method.

[0026] Figure 4 This is a schematic diagram of the operation interface of the first detection method.

[0027] Figure 5 It is a schematic diagram of highlighting the target waveform of the local online signal in the first detection mode.

[0028] Figure 6 It is a schematic diagram of highlighting the target waveform in the first detection mode by distributing the global online signal in an electrode sequence.

[0029] Figure 7 It is a schematic diagram of highlighting the target waveform in the left and right brain distribution of the global online signal in the first detection mode.

[0030] Figure 8 It is a schematic diagram of the operation interface of waveform clustering in the first detection method.

[0031] Fig. 9 It is a schematic diagram of the waveform clustering results of channel O2-Av in the first detection method.

[0032] Fig.10 It is a schematic diagram of the result after channel O2-Av screens the signal template in the first detection mode.

[0033] Fig.11 This is a schematic diagram of the operation interface for adjusting the matching degree in the first detection method.

[0034] Fig.12 It is a schematic diagram of highlighting the target waveform after adjusting the matching degree in the first detection method.

[0035] Fig.13 This is the workflow diagram of the second detection method.

[0036] Fig.14 It is a schematic diagram of the operation interface for online selection of the target signal as the second signal template in the second detection mode.

[0037] Fig.15 It is a schematic diagram of the operation interface for local pre-adjustment of matching degree in the second detection method.

[0038] Fig.16 It is a schematic diagram of highlighting the target waveform of a single-channel local online signal in the second detection method.

[0039] Fig.17 It is a schematic diagram of highlighting the target waveform of the multi-channel local online signal in the second detection mode.

[0040] Fig.18 It is a schematic diagram of the operation interface of the global search in the second detection method.

[0041] Fig.19 It is a schematic diagram of highlighting the target waveform in the second detection mode by distributing the global online signal in an electrode sequence.

[0042] Fig. 20 This is a workflow diagram of the third detection method.

[0043] Fig.21 It is a schematic diagram of the operation interface of discharge frequency analysis in quantitative analysis.

[0044] Fig. 22 It is the histogram of discharge frequency analysis in quantitative analysis.

[0045] Fig.23 It is a 3D brain topography map for discharge frequency analysis in quantitative analysis.

[0046] Fig.24 This is a schematic diagram of the operation interface of intensity analysis in quantitative analysis.

[0047] Fig.25 It is a 3D brain topography map and time-frequency map for intensity analysis in quantitative analysis. DETAILED DESCRIPTION

[0048] 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.

[0049] Example 1.

[0050] See Figure 1 , this embodiment 1 provides an online detection method for a target signal, including: signal detection to obtain a target signal set and its corresponding signal template; signal extraction, i.e. extracting the target signal using the waveform characteristics of the online signal; iterative optimization, i.e. adjusting the matching threshold between the target signal and the signal template according to the extraction result of the target signal to screen the target signal and update the target signal set.

[0051] Optionally, the online signal includes but is not limited to physiological electrical signals, such as electroencephalogram signals, and the target signal may be a selected typical signal waveform or an epileptic discharge waveform for detecting epileptic seizures. The target waveform or characteristic parameters of the target signal may be manually selected by using the signal waveform diagram of the online signal to extract the target waveform online.

[0052] Optionally, the matching degree between the target signal and the signal template is, for example but not limited to, waveform similarity.

[0053] Optionally, the signal extraction includes: obtaining the waveform position of the target signal, that is, selecting the start point and end point of the target signal in the waveform display diagram; obtaining the channel to which the target signal belongs; and highlighting the target signal through the signal waveform diagram of the online signal, such as but not limited to distinguishing the target signal in the signal waveform diagram by color, brightness, etc., so as to facilitate the extraction of results by manual observation and intervene to adjust the matching degree and characteristic parameters, etc., to achieve iterative optimization of the target signal set. Optionally, the highlighting includes, for example but not limited to, highlighting, color differentiation, bolding, changing line type display, etc., so that the operator can observe the extraction results through the operation interface.

[0054] Furthermore, in this case, there can be multiple signal detection methods to obtain the target signal set and its corresponding signal template, including: any one of the first detection method, the second detection method, and the third detection method. In order to clearly describe the differences between various detection methods and their applicability, the specific operation process, namely, signal detection, signal extraction, iterative optimization, etc., is explained in conjunction with the operation interface of the human-computer interaction machine of the epilepsy signal detection system. Figure 2 Taking 19-channel EEG data as an example, function options are set above the operation interface of the human-computer interaction machine, and the area below it is respectively provided with a local display area and a global display area; the local display area is used to display the signal waveform of the local online signal, and the end of the signal waveform corresponds to the channel number of the online signal, such as from P1-Av to Oz-Av, which also represents the electrode position of the online signal, and its time period width can be set from 170-180s or other time periods, which is convenient for extracting the target signal and adjusting parameters and iterative optimization; the global display area is used to display the signal waveform of the global online signal, and its time period width can be set from 0-900s or other time periods, which is convenient for displaying the target signal in the global online signal.

[0055] As a first optional implementation method of signal detection.

[0056] See Figure 3 , the first detection mode is configured as automatic detection, including: initially setting characteristic parameters of the target signal; local pre-adjustment, that is, adjusting the characteristic parameters in the local online signal to change the extraction result of the target signal; global search, that is, detecting the target signal according to the characteristic parameters in the global online signal to obtain a first target signal set; performing waveform clustering on the first target signal set to obtain the first signal template. Figure 4-Figure 12 ,The specific process of the online detection method using the first ,detection method is as follows.

[0057] (1) After selecting the channel and time limit, initially set the characteristic parameters of the target signal. Figure 4, select the "Automatic Detection" option in the function options on the operation interface of the human-computer interaction machine. The drop-down menu options include "01-High Sensitivity Detection" option, "02-Waveform Clustering" option, "03-High Specificity Detection" option, and "Cancel Search" option; select the "01-High Sensitivity Detection" option, and then the "Local Pre-adjustment Parameters" option, "Global Detection" option, and "Cancel High Sensitivity Detection" option will pop up; select the "Local Pre-adjustment Parameters" option to perform local pre-adjustment, and pop up feature parameters, including but not limited to selecting time period (s), amplitude threshold (uV), energy threshold (uV), amplitude time limit (s), and energy time limit (s).

[0058] (2) Use the signal waveform of the online signal to extract the target signal from the local online signal, so that the target signal is highlighted in the signal waveform (such as Figure 5 The target waveform is highlighted in the local display area, for example, the target waveform is highlighted between 177s and 178s in channel P3-Av). By adjusting the feature parameters and observing whether the extraction result (i.e., the display result) of the target signal meets the set waveform extraction requirements, try to ensure that all the required target waveforms are covered. Obviously, by setting the feature parameters, more target waveforms can be obtained in the local online signal, and more clustering categories can be obtained after global retrieval.

[0059] (3) Select the "Global Detection" option to perform global search. Search for target signals in the global online signal based on the characteristic parameters to obtain the first target signal set for preliminary screening; the horizontal axis is time, the vertical axis is channel, and red represents the extracted waveform position. Then adjust the electrode distribution form to distribute the electrodes in order (such as Figure 6 ) or left-right brain distribution (e.g. Figure 7 The extraction results of the global online signal are displayed in the local display area and the global display area respectively in the form of the target waveform highlighted in the figure.

[0060] (4) Select the "02-Waveform Clustering" option to perform waveform clustering on the first target signal set. For example, select the "Sequential Distribution" option and perform waveform clustering (such as Figure 8 As shown in the figure), click the channel number on the far left of the signal waveform, and a pop-up window will display the clustering result of the current channel. Taking channel O2-Av as an example, its clustering result is as follows Fig. 9 As shown in the figure, a total of 4 cluster categories are displayed, and the signal templates in each cluster category are highlighted in the middle area of ​​the cluster signal. Obviously, cluster category 3 belongs to the noise signal. By checking the "Delete" checkbox and clicking the "Confirm Delete" button, the checked cluster category is deleted to obtain the first signal template corresponding to each category in the channel, that is, Fig.10 The signal templates corresponding to cluster category 1, cluster category 2, and cluster category 4 in .

[0061] (5) Select the "03-High Specificity Detection" option for template matching. In the menu below, the "Similarity Adjustment" option and the "Undo High Specificity Detection" option pop up; select the "Similarity Adjustment" option, and a pop-up window will display the waveform similarity adjustment interface (such as Fig.11 ), set the channel to be processed and the waveform similarity threshold. Remove the waveform containing noise according to the similarity threshold and update the online signal waveform graph, where the number of target signals changes (such as Fig.12 ), the similarity threshold can be repeatedly adjusted according to the extraction result, the unmatched target signals are deleted and the first target signal set is updated, the first signal template is updated according to the mean of the remaining target signals in the updated first target signal set, and the above operations are repeated until the matching degree of all target signals in the first target signal set with the updated signal template is greater than the matching threshold; the matching threshold is adjusted based on the extraction result of the target signal, the template matching is repeated, and the iterative optimization is performed until the target signal extraction result meets the expectations, at which time the corresponding first target signal set can be used as a prerequisite for quantitative analysis.

[0062] The first detection method of this embodiment is an automatic detection method of online signals. The first signal template obtained by detecting the target signal online based on the preset characteristic parameters is implemented in the form of actively fitting the signal template, which ensures the comprehensiveness of the signal template. First, the characteristic parameters of the target waveform are set in the local online signal through the signal waveform diagram, and then the target signal is retrieved in the global online signal to obtain its corresponding data set, and the signal template is fitted according to the target signal set. It is generally suitable for operators who cannot accurately find the target waveform through the signal waveform diagram or for the operation requirements that need to fully cover a variety of typical waveforms. Although the first signal template is obtained by clustering the first target signal set, generally each channel has multiple first signal templates, which can ensure the comprehensive coverage of the target waveform. Compared with the historical signal, the first signal template obtained by clustering is closer to the actual online signal, but there is still a certain difference between the two. This difference will cause the detection result of the online signal to have a certain error. Therefore, the matching threshold between the target signal and the signal template can be adjusted according to the extraction result of the target signal, and iterative optimization is performed through template matching to delete the noise signal or filter the target signal through the signal template, and update the first target signal set and its corresponding first signal template. The signal template is optimized by utilizing the data distribution characteristics of the target signal in the online signal, thereby improving the detection accuracy of the target signal.

[0063] As a second optional implementation manner of signal detection.

[0064] See Fig.13, the second detection mode is configured as semi-automatic detection, including: obtaining a second signal template, that is, using the signal waveform diagram of the online signal to extract the target waveform as the second signal template; local pre-adjustment of the matching degree, that is, adjusting the matching degree threshold in the local online signal to change the extraction result of the target signal; global search, that is, searching the target signal according to the matching degree threshold in the global online signal to obtain the second target signal set. Figure 14-Figure 19 ,The specific process of the online detection method using the second ,detection method is as follows.

[0065] (1) Extract the waveform online as the second signal template. Fig.14 , select the "Semi-automatic detection" option in the function options on the operation interface of the human-computer interaction machine. The drop-down menu options include "01-Template selection" option, "02-Local pre-adjustment parameter" option, "03-Global search" option, and "Undo search" option; select the "01-Template selection" option, then click the required channel number, click the signal waveform diagram of the online signal to select the required start point and end point, extract the target waveform as the template waveform and highlight it, such as Fig.14 The middle channel F4-Av shows a target waveform highlighted between 177s and 178s.

[0066] (2) Perform local pre-adjustment of matching degree. Fig.15 , select the "02-Local Pre-adjustment Parameters" option, and the matching parameters will pop up, including similarity parameters (between 0.1-1, 1 means the waveform of the target signal is the same as that of the signal template) and time range. Fig.16 , first select a single channel to set the similarity parameter threshold, and then highlight the extracted waveform, such as Fig.16 The middle channel F4-Av shows two target waveforms at 171s-172s and 177s-178s respectively. Fig.17 , you can extract signal templates and set similarity parameter thresholds for multiple channels one by one, and highlight the extracted target waveform in the figure. The similarity threshold parameters set for each channel can be inconsistent to extract different target waveforms, such as Fig.17The four target waveforms highlighted on the middle channel P2-Av are 170s-171s, 171s-172s, 174s-175s, and 177s-178s, respectively. The two target waveforms highlighted on the channel F4-Av are 171s-172s and 177s-178s, respectively. The four target waveforms highlighted on the F7-Av channel are 171s-172s, 173s-174s, and 177s -178s, there is 1 target waveform each; the F8-Av channel highlights 6 target waveforms, namely, 2 target waveforms between 170s-171s, 1 target waveform between 171s-172s, 2 target waveforms between 173s-174s, and 1 target waveform between 177s-178s; the Cz-Av channel highlights 4 target waveforms, namely, 3 target waveforms between 174s-175s, and 1 target waveform between 178s-179s.

[0067] (3) Select the "03-Global Search" option to perform a global search. Fig.18 , in the global online signal, all channels are extracted with the set matching degree, the extracted target waveform is highlighted, and the global extracted waveform is displayed at the bottom of the operation interface of the human-computer interaction machine (see Fig.19 The target waveform displayed in the global display area) is the second target signal set. Fig.19 and Fig.12 From the local display area and the global display area, it can be clearly seen that the target waveform data obtained by setting the characteristic parameters to obtain the template signal (automatic detection) is more comprehensive than the target template actively selected (semi-automatic detection). Of course, this difference depends more on the operator's expertise in identifying the target signal.

[0068] (4) Adjust the matching threshold. If the extraction result does not meet the requirements, you can reselect the second signal template and perform a global search to update the second target signal set, or perform a local pre-adjustment of the matching degree to update the second target signal set. Iteratively optimize until the extraction result of the target signal meets the set extraction requirements. At this time, the corresponding second target signal set can be used as a prerequisite for quantitative analysis.

[0069] The second detection method of this embodiment is configured as semi-automatic detection, which is to detect the target signal online based on the second signal template selected by the signal waveform, and is implemented by actively selecting the signal template to ensure the accuracy of the signal template. The target waveform is first accurately found through the signal waveform diagram, and then the target waveform is used as a signal template to retrieve the target signal set. In the iterative optimization, the waveform similarity threshold between the target signal and the signal template is adjusted according to the extraction result of the target signal to screen the target signal. During the operation, the signal template can be kept unchanged and the second target signal set can be updated. It is generally suitable for experienced operators or the operation requirements of accurately locking the target waveform. Although the comprehensiveness of the target signal needs to rely on the professional experience of the operator, using the target signal in the online signal as the signal template eliminates the difference between the signal template and the online signal, which can maximize the accuracy of the signal detection result.

[0070] As a third optional implementation of signal detection.

[0071] See Fig. 20 , the third detection method is configured as a combination of automatic detection and semi-automatic detection, including: waveform merging the target signal sets obtained by the first detection method and the second detection method to obtain a third target signal set. Specifically, after the waveforms are merged, the third signal template obtained by waveform clustering includes both the first signal template obtained by clustering fitting and the second signal template obtained by the actual online signal, that is, when the first signal template is the same as the second signal template, the same template can be eliminated by waveform clustering, and the target signal set obtained by waveform retrieval is not affected. When the first signal template is different from the second signal template, the target signal set obtained by waveform retrieval can be expanded, and a new signal template can be obtained again by waveform clustering, and then the signal template of the clustering result is manually screened or deleted to remove inappropriate clustering categories to ensure the comprehensiveness and accuracy of the target signal. The acquisition method of the first target signal set and its corresponding first signal template, the second target signal set and its corresponding second signal template is referred to the above content, which will not be repeated here. When the signal detection is configured as the third detection method, the iterative optimization is configured as template matching, and its specific process is referred to the relevant description in the first detection method.

[0072] Example 2.

[0073] On the basis of Example 1, this Example 2 provides an epilepsy signal detection system, including: a processor, running the online detection method; a human-computer interaction machine connected to the processor, and its operation interface is provided with function options, a signal extraction area, and quantitative analysis options; wherein the function options are provided with a signal detection method option box to select different methods for signal detection; the signal extraction area is suitable for displaying a signal waveform diagram of the online signal for extracting the target signal; the quantitative analysis option is suitable for quantitatively analyzing the target signal in the target signal set to obtain a measurement indicator of the epileptic discharge signal.

[0074] Furthermore, the quantitative analysis options include at least one of discharge frequency analysis, intensity analysis, and spatial time series analysis. The target signals extracted after automatic detection, semi-automatic detection, or a combination of the two can all be quantitatively analyzed. Now, taking the first data set obtained by automatic detection as an example, the operation process of each quantitative analysis method is described as follows:

[0075] Optional, see Fig.21 The discharge frequency analysis includes: selecting the required channel, counting the discharge frequency of the target signal, calculating the frequency of global or certain type of discharge in each channel, and drawing a histogram of the corresponding discharge frequency of all channels (such as Fig. 22 ) and 3D brain topography (as shown Fig.23 As shown, the frequency intensity is represented by gradient colors).

[0076] Optional, see Fig.24 The intensity analysis includes: drawing the time-frequency-space energy spectrum of the target signal, selecting the display period, interval (i.e., window length), frequency range, and display channel; drawing the 3D brain topography map of the waveform energy extracted in each window and the time-frequency map of the display period; adjusting the time to display the waveform energy changes of the target signal at different times through gradient colors. The change of waveform energy in the time period with the interval as the window at the current display moment, Fig.25 This is a time-frequency diagram for the complete time period, corresponding to the slider time, and the corresponding color range in the diagram can be adjusted. At 343 seconds, it can be seen in the time-frequency diagram that there is an obvious increase in high-frequency and low-frequency energy. Corresponding to the topographic map, it is found that there is an obvious discharge signal in the forehead.

[0077] Optionally, the spatial timing analysis includes: taking the target signal set as a reference, intercepting the signal data of all channels in the same time period, and the signal data of each time period is a group; dividing the groups of waveforms with the same channel position corresponding to the maximum waveform into one category, and calculating the mean of the waveforms of each channel in each category as a signal template for display; determining the peak moment corresponding to the maximum waveform of each type of signal template, and drawing a 3D brain topography map with the voltage value of each channel at the current moment.

[0078] Obviously, this epilepsy signal detection system can select different detection methods according to the operator's professionalism and detection requirements through the function options of the operation interface, and obtain the measurement indicators of the epileptic discharge signal through the signal extraction area and quantitative analysis options. It can not only obtain a more comprehensive target signal set, but also improve the accuracy of the signal template and the target signal set. Epilepsy detection can be achieved using only online signals. At the same time, the personal expertise of the operator is combined with the statistical analysis function of the processor, which is not only easy to operate, but also reduces the universal professional requirements of the equipment and expands the application group of the equipment.

[0079] Example 3.

[0080] Based on the first embodiment, the third embodiment 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.

[0081] Example 4.

[0082] Based on Example 1, this Example 4 provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0083] Example 5.

[0084] Based on Embodiment 1, this Embodiment 5 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.

[0085] 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.

[0086] 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.

[0087] 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.

[0088] With the above-mentioned ideal embodiments of the present invention as inspiration, through the above-mentioned description content, relevant personnel can make various changes and modifications without departing from the scope of the technical idea of ​​the present invention.

Claims

1. An online detection method for a target signal, characterized in that: include: Signal detection to obtain the target signal set and its corresponding signal template; Signal extraction, i.e. using the waveform characteristics of online signals to extract target signals; Iterative optimization, i.e. adjusting the matching threshold between the target signal and the signal template according to the target signal extraction result to screen the target signal and update the target signal set; The signal detection includes any one of the first detection mode and the third detection mode; in The first detection mode is configured as a first signal template obtained based on online detection of a target signal; The third detection mode is configured to obtain a target signal by combining the first detection mode and the second detection mode; The second detection mode is configured to detect the target signal online based on a second signal template selected based on the signal waveform; The first detection method includes: Initially set the characteristic parameters of the target signal; Local pre-adjustment of parameters, i.e., adjusting the characteristic parameters in the local online signal to change the extraction result of the target signal; Global search, i.e. detecting target signals in global online signals according to characteristic parameters to obtain a first target signal set; Perform waveform clustering on the first target signal set to obtain the cluster category of the current channel, highlight the signal templates in each cluster category in the middle area of ​​the cluster signal, delete the cluster category belonging to the noise signal by checking the "Delete" checkbox, and obtain a waveform template as the first signal template; The characteristic parameters include: time period, amplitude threshold, energy threshold, amplitude time limit, and energy time limit.

2. The online detection method according to claim 1, characterized in that: When the signal detection is configured as the first detection mode, the iterative optimization is configured as template matching, including: Detect the matching degree between the target signal in the target signal set and the signal template, remove the target signal with the lowest matching degree and below the matching degree threshold in the target signal set, update the target signal set and update the signal template with the average value of the remaining target signals, and repeat the above operation until the matching degree between all the target signals in the target signal set and the updated signal template is greater than the matching degree threshold; The matching degree threshold is adjusted based on the target signal extraction result, and the template matching is repeated until the target signal extraction result meets expectations.

3. The online detection method according to claim 1, characterized in that: The second detection method includes: Acquire a second signal template, that is, extract a target waveform as the second signal template using a signal waveform diagram of an online signal; Locally pre-adjusting the matching degree, that is, adjusting the matching degree threshold in the local online signal to change the extraction result of the target signal; Global search, that is, searching for target signals in global online signals according to a matching degree threshold to obtain a second target signal set.

4. The online detection method according to claim 1 or 3, characterized in that: When the signal detection is configured as the second detection mode, the iterative optimization is configured to adjust the matching degree threshold, including: The matching degree threshold is adjusted based on the extraction result of the target signal, and the template matching is repeated until all target signals in the target signal set meet the set extraction result of the target signal.

5. The online detection method according to claim 1, characterized in that: The third detection method includes: Merging the waveforms of the target signal sets obtained by the first detection method and the second detection method to obtain a third target signal set; Waveform clustering is performed on the third target signal set to obtain a third signal template.

6. The online detection method according to claim 1, characterized in that: The signal extraction comprises: Get the waveform position of the target signal, that is, select the start point and end point of the target signal in the waveform display; Obtain the channel to which the target signal belongs; and Highlight the target signal through the signal waveform diagram of the online signal.

7. An epilepsy signal detection system, characterized in that: include: A processor, executing the online detection method according to claim 1; A human-computer interaction machine connected to the processor, whose operation interface is provided with function options, signal extraction area, and quantitative analysis options; in The function options include a signal detection mode option box to select different modes for signal detection; The signal extraction area is suitable for displaying a signal waveform diagram of an online signal for extracting a target signal; The quantitative analysis option is suitable for performing quantitative analysis on the target signal in the target signal set to obtain a measurement index of the epileptic signal.

8. The signal detection system according to claim 7, characterized in that: The quantitative analysis options include at least one of discharge frequency analysis, intensity analysis, and spatial time series analysis; in The discharge frequency analysis includes: selecting the required channel, counting the discharge frequency of the target signal, calculating the frequency of global or certain type of discharge in each channel, and drawing a histogram and a 3D brain topography of the corresponding discharge frequencies of all channels; The intensity analysis includes: drawing a time-frequency-space domain energy spectrum for the target signal, selecting a display period, interval, frequency range, and display channel; drawing a 3D brain topography map of the extracted waveform energy in each window and a time-frequency map of the display period; adjusting the time to display the waveform energy changes of the target signal at different times; The spatial time series analysis includes: taking the target signal set as a reference, intercepting the signal data of all channels in the same time period, and the signal data of each time period is a group; classifying the groups of waveforms with the same channel position corresponding to the maximum waveform into one category, calculating the mean value of the waveform of each channel in each category as a signal template for display; determining the peak moment corresponding to the maximum waveform of each type of signal template, and drawing a 3D brain topography map based on the voltage value of each channel at the current moment.

9. 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 claim 1.

10. 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 claim 1 are implemented.

11. 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 claim 1 are implemented.

Images