Medical metrology detection device and method

Abstract

The invention discloses a medical measurement detection device and method, and the method comprises the steps: collecting an electrocardiogram signal of a detected patient, and introducing a signal processing and analysis algorithm based on artificial intelligence and deep learning at a rear end, so as to carry out the analysis of the electrocardiogram signal; a CEEMDAN decomposition technology is used for decomposing an electrocardiogram signal to extract an intrinsic mode function (IMF) component signal of the electrocardiogram signal, and then an artificial intelligence technology is used for learning and depicting full-frequency-domain association and aggregation feature representation between waveform semantics of the frequency band IMF component signal related to the electrocardiogram signal; therefore, whether the heart rate of the detected patient is abnormal or not can be detected. Therefore, the electrocardiogram signal can be automatically analyzed by utilizing an artificial intelligence technology, so that more intelligent medical measurement detection is realized, and in this way, human intervention can be effectively reduced, and the problems of misdiagnosis, missed diagnosis and low efficiency caused by a traditional mode are avoided.

Description

Technical Field

[0001] The present application relates to the field of intelligent detection, and more specifically, to a medical metrology detection device and method. Background Art

[0002] Electrocardiogram is a non-invasive examination method widely used in clinical medicine. It evaluates the functional state and health of the heart by measuring and detecting changes in the heart's electrical activity. Arrhythmia is a common heart disease, and timely and accurate detection is crucial to the treatment and prognosis of patients. However, traditional electrocardiogram analysis usually relies on the doctor's experience and interpretation, which is easily affected by subjective factors, and different doctors may give different conclusions. At the same time, it also ignores the complex patterns and subtle differences hidden in the signal, and cannot accurately detect mild or intermittent arrhythmias, resulting in misdiagnosis or missed diagnosis. In addition, the manual analysis of electrocardiograms has a low degree of intelligence and takes a lot of time, especially when faced with large amounts of long-term electrocardiogram monitoring data, which is inefficient. In addition, due to the complexity of the electrocardiogram waveform, it becomes more difficult to identify specific arrhythmias, especially when other heart diseases are present.

[0003] Therefore, an optimized medical measurement detection device is desired. Summary of the invention

[0004] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a medical metrology detection device and method, which collects the electrocardiogram signal of the patient being detected, and introduces a signal processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the electrocardiogram signal, so as to decompose the electrocardiogram signal using the CEEMDAN decomposition technology to extract its intrinsic mode function (IMF) component signal, and then use artificial intelligence technology to learn and characterize the full frequency domain association and aggregation feature representation between the waveform semantics of the frequency band IMF component signal of the electrocardiogram signal, so as to detect whether the heart rate of the patient being detected is abnormal. In this way, artificial intelligence technology can be used to automatically analyze the electrocardiogram signal, thereby realizing more intelligent medical metrology detection. In this way, human intervention can be effectively reduced, and the problems of misdiagnosis, missed diagnosis and low efficiency caused by traditional methods can be avoided.

[0005] According to one aspect of the present application, a medical measurement detection device is provided, comprising: An electrocardiogram signal acquisition module, used for acquiring electrocardiogram signals; An electrocardiogram signal decomposition module, used for performing CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; An electrocardiogram component signal waveform feature extraction module, used for respectively extracting features from each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; An electrocardiogram component signal waveform feature association optimization module, used for inputting the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; A full-frequency domain electrocardiogram signal waveform semantic aggregation representation module is used to aggregate the set of the optimized electrocardiogram frequency band IMF component signal waveform feature vectors to obtain the full-frequency domain electrocardiogram signal waveform significant semantic features; The arrhythmia detection module is used to determine a measurement detection result based on the significant waveform semantic features of the full-frequency domain electrocardiogram signal, and the measurement detection result is used to indicate whether an arrhythmia exists.

[0006] According to another aspect of the present application, a medical metrology detection method is provided, which includes: Acquire electrocardiogram signals; Performing CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; Performing feature extraction on each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals respectively to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; Inputting the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; Aggregating the set of waveform feature vectors of the optimized electrocardiogram frequency band IMF component signal to obtain significant waveform semantic features of the full-frequency domain electrocardiogram signal; Based on the significant waveform semantic features of the full-frequency domain electrocardiogram signal, a measurement detection result is determined, and the measurement detection result is used to indicate whether there is an arrhythmia.

[0007] Compared with the prior art, the medical metrology detection equipment and method provided by the present application collects the electrocardiogram signal of the patient being detected, and introduces a signal processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the electrocardiogram signal, so as to decompose the electrocardiogram signal using the CEEMDAN decomposition technology to extract its intrinsic mode function (IMF) component signal, and then use artificial intelligence technology to learn and characterize the full frequency domain association and aggregation feature representation between the waveform semantics of the frequency band IMF component signal of the electrocardiogram signal, so as to detect whether the heart rate of the patient being detected is abnormal. In this way, artificial intelligence technology can be used to automatically analyze the electrocardiogram signal, thereby realizing more intelligent medical metrology detection. In this way, human intervention can be effectively reduced, and the problems of misdiagnosis, missed diagnosis and low efficiency caused by traditional methods can be avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 is a block diagram of a medical metrology detection device according to an embodiment of the present application; Figure 2 A schematic diagram of data flow of a medical measurement and detection device according to an embodiment of the present application; Figure 3 Flow chart of a medical metrology detection method according to an embodiment of the present application. DETAILED DESCRIPTION

[0010] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0011] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0012] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0013] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0014] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0015] Traditional ECG analysis usually relies on the doctor's experience and interpretation, which is easily affected by subjective factors, and different doctors may come to different conclusions. At the same time, it also ignores the complex patterns and subtle differences hidden in the signal, and cannot accurately detect mild or intermittent arrhythmias, resulting in misdiagnosis or missed diagnosis. In addition, the manual analysis of ECGs is less intelligent and takes a lot of time, especially when faced with large amounts of long-term ECG monitoring data, which is inefficient. In addition, due to the complexity of the ECG waveform, it becomes more difficult to identify specific arrhythmias, especially when other heart diseases are present. Therefore, an optimized medical metrology detection device is desired.

[0016] In the technical solution of the present application, a medical metrology detection device is proposed. Figure 1 is a block diagram of a medical metrology detection device according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of medical measurement and detection equipment according to an embodiment of the present application. Figure 1 and Figure 2As shown, the medical metrology detection device 300 according to the embodiment of the present application includes: an electrocardiogram signal acquisition module 310, used to acquire an electrocardiogram signal; an electrocardiogram signal decomposition module 320, used to perform CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; an electrocardiogram component signal waveform feature extraction module 330, used to perform feature extraction on each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; an electrocardiogram component signal waveform feature association optimization module 340, used to extract the electrocardiogram IMF component signal waveform feature vectors from the electrocardiogram IMF component signal waveform feature vectors; A set of ECG frequency band IMF component signal waveform feature vectors is input into a feature association optimization network based on a semantic measure of sequence endogenous correlation to obtain a set of optimized ECG frequency band IMF component signal waveform feature vectors; a full-frequency domain ECG signal waveform semantic aggregation representation module 350 is used to aggregate the set of optimized ECG frequency band IMF component signal waveform feature vectors to obtain full-frequency domain ECG signal significant waveform semantic features; an arrhythmia detection module 360 ​​is used to determine a measurement detection result based on the full-frequency domain ECG signal significant waveform semantic features, and the measurement detection result is used to indicate whether there is an arrhythmia.

[0017] In particular, the electrocardiogram signal acquisition module 310 is used to acquire an electrocardiogram signal, wherein the electrocardiogram can record the electrical activity of the heart in real time and provide information on the heart rhythm, heart rate and the electrical conduction state of the heart.

[0018] In particular, the ECG signal decomposition module 320 is used to perform CEEMDAN decomposition on the ECG signal to obtain a set of ECG IMF component signals. It is worth mentioning that CEEMDAN is a signal processing technology used to decompose a complex signal into multiple simpler intrinsic mode functions (IMF) to obtain a set of ECG IMF component signals. This method is very effective in nonlinear and non-stationary signal processing. By performing CEEMDAN decomposition on the ECG signal, the ECG signal characteristics of different frequency ranges and different frequency bands can be captured to facilitate the analysis and understanding of the patient's cardiac electrophysiological activities. For example, some IMF components may be associated with specific physiological states of the heart (such as arrhythmia), which is crucial to improving the accuracy of subsequent analysis.

[0019] In particular, the ECG component signal waveform feature extraction module 330 is used to perform feature extraction on each ECG IMF component signal in the set of ECG IMF component signals to obtain a set of ECG frequency band IMF component signal waveform feature vectors. In a specific example of the present application, each ECG IMF component signal in the set of ECG IMF component signals is input into a component signal feature extractor based on a one-dimensional convolutional neural network model to obtain a set of ECG frequency band IMF component signal waveform feature vectors. It should be understood that although CEEMDAN can decompose an ECG signal into multiple ECG IMF component signals, these ECG IMF component signals themselves do not directly provide sufficient information to distinguish between normal and abnormal ECG patterns, nor can they identify whether a patient has arrhythmia. Therefore, in the technical solution of the present application, each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals is further input into a component signal feature extractor based on a one-dimensional convolutional neural network model for feature mining, so as to extract the waveform implicit feature information of each electrocardiogram IMF component signal, thereby obtaining a set of electrocardiogram frequency band IMF component signal waveform feature vectors. It is worth mentioning that a one-dimensional convolutional neural network (1D CNN) is a neural network model specifically used to process one-dimensional data. It is similar to a traditional two-dimensional convolutional neural network (2D CNN), but its convolution kernel slides on the one-dimensional space of the input data instead of on the two-dimensional space.

[0020] In particular, the electrocardiogram component signal waveform feature association optimization module 340 is used to input the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors. Since each electrocardiogram frequency band IMF component signal waveform feature vector in the set of electrocardiogram frequency band IMF component signal waveform feature vectors respectively represents the semantic features of the electrocardiogram IMF component signal waveform of the corresponding frequency range and frequency band, and there are interactive correlations and mutual influences between the semantic features of the electrocardiogram IMF component signal waveforms of different frequency bands, which jointly play a role in the subsequent arrhythmia detection task. However, some of the semantic features of the electrocardiogram IMF component signal waveforms of different frequency bands are crucial to the subsequent detection task, while some are irrelevant interference information or redundant information. Based on this, in order to be able to identify feature information with higher relevance to subsequent tasks from the set of feature vectors of the ECG frequency band IMF component signal waveform, thereby strengthening the semantic expression of these signal waveform feature components in the subsequent feature aggregation process and achieving more accurate arrhythmia detection tasks, in the technical solution of the present application, the set of feature vectors of the ECG frequency band IMF component signal waveform is further input into a feature association optimization network based on the semantic measurement of sequence endogenous correlation to obtain a set of optimized ECG frequency band IMF component signal waveform feature vectors. In particular, the feature association optimization network based on the semantic measurement of sequence endogenous correlation can improve the quality of feature representation by quantifying and optimizing the semantic associations between the feature vectors of the ECG frequency band IMF component signal waveform.

[0021] In an embodiment of the present application, the set of ECG frequency band IMF component signal waveform feature vectors is input into a feature association optimization network based on the semantic measurement of sequence endogenous correlation to obtain a set of optimized ECG frequency band IMF component signal waveform feature vectors, including: first, calculating the semantic association score vector between any two ECG frequency band IMF component signal waveform feature vectors in the set of ECG frequency band IMF component signal waveform feature vectors to obtain a set of ECG frequency band IMF component signal waveform semantic association score vectors; thereby quantifying the similarity or difference between the IMF component signal waveform features of different ECG frequency bands, providing a basis for understanding the semantic structure within the sequence. Next, based on the set of ECG frequency band IMF component signal waveform semantic association score vectors, determine the global representation vector of the endogenous correlation of the ECG frequency band IMF component signal waveform sequence; in this way, the semantic structure of the set of the entire ECG frequency band IMF component signal waveform feature vectors is captured, providing an overall context for subsequent association optimization. Then, based on the global representation vector of the endogenous correlation of the ECG frequency band IMF component signal waveform sequence, the correlation optimization factor of each ECG frequency band IMF component signal waveform feature vector in the set of the ECG frequency band IMF component signal waveform feature vectors is calculated to obtain a sequence of ECG frequency band IMF component signal waveform semantic correlation optimization factors; wherein the ECG frequency band IMF component signal waveform semantic correlation optimization factors highlight the consistency or deviation between the individual feature vector and the overall semantic structure of the sequence, and provide guidance for feature adjustment. Furthermore, the sequence of the ECG frequency band IMF component signal waveform semantic correlation optimization factors is input into Activation function is used to obtain a sequence of semantic association optimization weight factors of the ECG frequency band IMF component signal waveform; that is, in order to enhance the model's ability to distinguish the weights of the IMF component signal waveform features in different frequency bands, so that the weights are more suitable for feature optimization and subsequent arrhythmia detection tasks, the association optimization factor sequence is further input into the Softmax activation function to be converted into weight factors to achieve factor normalization and nonlinear mapping. Subsequently, each ECG frequency band IMF component signal waveform semantic association optimization weight factor in the sequence of ECG frequency band IMF component signal waveform semantic association optimization weight factors is used as a weight to weight each ECG frequency band IMF component signal waveform feature vector in the set of ECG frequency band IMF component signal waveform feature vectors to obtain the set of optimized ECG frequency band IMF component signal waveform feature vectors. Here, the IMF component signal waveform feature information and optimization factors of each frequency band of the original ECG are combined through weighted operation to enhance the expressiveness of each ECG frequency band IMF component signal waveform feature, reduce the interference of irrelevant or noise features, and strengthen the contribution of key features. This feature optimization method based on the semantic measurement of endogenous correlation of the sequence is particularly useful in processing the semantic features of IMF component signal waveforms in different frequency bands in complex ECG signals, because it can help the model better understand the subtle changes and patterns in the signal.

[0022] Among them, the process of calculating the semantic association score vector between any two ECG frequency band IMF component signal waveform feature vectors in the set of ECG frequency band IMF component signal waveform feature vectors to obtain a set of ECG frequency band IMF component signal waveform semantic association score vectors includes: cascading any two ECG frequency band IMF component signal waveform feature vectors in the set of ECG frequency band IMF component signal waveform feature vectors to obtain a set of ECG frequency band IMF component signal waveform semantic cascade feature vectors; multiplying each ECG frequency band IMF component signal waveform semantic cascade feature vector in the set of ECG frequency band IMF component signal waveform semantic cascade feature vectors with the corresponding ECG frequency band IMF component weight matrix, and then adding them with the ECG frequency band IMF component bias vector by position to obtain the set of ECG frequency band IMF component signal waveform semantic association score vectors.

[0023] More specifically, based on the set of semantic association score vectors of the ECG frequency band IMF component signal waveforms, the process of determining the global representation vector of the intrinsic correlation of the ECG frequency band IMF component signal waveform sequence includes: calculating the mean vector of the set of semantic association score vectors of the ECG frequency band IMF component signal waveforms as the global representation vector of the intrinsic correlation of the ECG frequency band IMF component signal waveform sequence. And, based on the global representation vector of the endogenous correlation of the electrocardiogram frequency band IMF component signal waveform sequence, the process of calculating the association optimization factor of each electrocardiogram frequency band IMF component signal waveform feature vector in the set of the electrocardiogram frequency band IMF component signal waveform feature vectors to obtain a sequence of semantic association optimization factors of the electrocardiogram frequency band IMF component signal waveform comprises: multiplying the global representation vector of the endogenous correlation of the electrocardiogram frequency band IMF component signal waveform sequence with the global weight matrix of the electrocardiogram frequency band IMF component signal waveform to obtain the global representation weight vector of the electrocardiogram frequency band IMF component signal waveform; respectively comparing each electrocardiogram frequency band IMF component signal waveform feature vector in the set of the electrocardiogram frequency band IMF component signal waveform feature vectors with the corresponding electrocardiogram frequency band IMF component signal waveform global representation weight vector; The local weight matrices of the electrocardiogram frequency band IMF component signal waveform are multiplied to obtain a set of local representation weight vectors of the electrocardiogram frequency band IMF component signal waveform; each local representation weight vector of the electrocardiogram frequency band IMF component signal waveform in the set of local representation weight vectors of the electrocardiogram frequency band IMF component signal waveform is calculated, and the local representation weight vector of each electrocardiogram frequency band IMF component signal waveform is respectively added by position with the global representation weight vector of the electrocardiogram frequency band IMF component signal waveform and the corresponding electrocardiogram frequency band IMF component signal waveform bias vector to obtain a set of semantic association vectors of the electrocardiogram frequency band IMF component signal waveform; the set of semantic association vectors of the electrocardiogram frequency band IMF component signal waveform is input into a sigmoid function to obtain a sequence of semantic association optimization factors of the electrocardiogram frequency band IMF component signal waveform.

[0024] In summary, in the above embodiment, the set of the electrocardiogram frequency band IMF component signal waveform feature vectors is input into the feature association optimization network based on the sequence endogenous correlation semantic measurement to obtain the set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors, including: inputting the set of the electrocardiogram frequency band IMF component signal waveform feature vectors into the feature association optimization network based on the sequence endogenous correlation semantic measurement and processing them with the following feature association optimization formula to obtain the set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; wherein the feature association optimization is: in, is the set of waveform feature vectors of the IMF component signal in the electrocardiogram frequency band, represents the number of feature vectors in the set of feature vectors of the IMF component signal waveform of the electrocardiogram frequency band, and are respectively the first and The IMF component signal waveform feature vectors of the ECG frequency band, For cascade processing operations, is the weight matrix of IMF components in the ECG frequency band, is the IMF component bias vector of the ECG frequency band, is the first in the set of semantic association score vectors of the ECG frequency band IMF component signal waveform ECG frequency band IMF component signal waveform semantic association score vector, is the number of vectors in the set of semantic association score vectors of the IMF component signal waveform of the electrocardiogram frequency band, is the global representation vector of the endogenous correlation of the IMF component signal waveform sequence in the electrocardiogram frequency band, and are the global weight matrix of the IMF component signal waveform in the ECG frequency band and the local weight matrix of the IMF component signal waveform in the ECG frequency band, is the bias vector of the IMF component signal waveform in the ECG frequency band, yes function, yes function, The first in the sequence of weight factors for optimizing the semantic association of the IMF component signal waveform of the electrocardiogram frequency band The weight factor of the IMF component signal waveform semantic association optimization of the ECG frequency band is The first one in the set of the optimized electrocardiogram frequency band IMF component signal waveform feature vectors Optimize the IMF component signal waveform feature vector of the ECG frequency band.

[0025] In particular, the full-frequency domain ECG signal waveform semantic aggregation representation module 350 is used to aggregate the set of the optimized ECG frequency band IMF component signal waveform feature vectors to obtain the full-frequency domain ECG signal significant waveform semantic features. In a specific example of the present application, the set of the optimized ECG frequency band IMF component signal waveform feature vectors is cascaded to obtain the full-frequency domain ECG signal significant waveform feature vector as the full-frequency domain ECG signal significant waveform semantic features. It should be understood that since the set of the optimized ECG frequency band IMF component signal waveform feature vectors has undergone feature optimization, it can more accurately represent the important information of each frequency component in the ECG signal, and remove unnecessary noise and redundant information. Furthermore, the set of the optimized ECG frequency band IMF component signal waveform feature vectors is cascaded to construct a full-frequency domain ECG signal significant waveform feature vector containing all frequency component information, wherein the feature vector contains all important frequency component information, not just the information of a single ECG signal frequency band.

[0026] In particular, the arrhythmia detection module 360 ​​is used to determine the measurement detection result based on the significant waveform semantic features of the full-frequency domain electrocardiogram signal, and the measurement detection result is used to indicate whether there is an arrhythmia. In a specific example of the present application, the significant waveform feature vector of the full-frequency domain electrocardiogram signal is input into a measurement detection result generator based on a classifier to obtain a measurement detection result, and the measurement detection result is used to indicate whether there is an arrhythmia. That is, the key significant waveform representation information of the full-frequency domain electrocardiogram signal is used for classification processing to detect whether the heart rate of the patient being tested is abnormal, thereby obtaining a measurement detection result. In this way, artificial intelligence technology can be used to automatically analyze the electrocardiogram signal, thereby realizing a more intelligent medical measurement detection. In this way, human intervention can be effectively reduced, and the problems of misdiagnosis, missed diagnosis and low efficiency caused by traditional methods can be avoided.

[0027] In a preferred example, considering that each ECG frequency band IMF component signal waveform feature vector in the set of optimized ECG frequency band IMF component signal waveform feature vectors represents the inter-domain optimized correlation feature of the corresponding ECG IMF component signal based on the intra-domain correlation measurement of the one-dimensional time series correlation feature in the local sample domain, the imbalance of the distribution of intra-domain time series correlation features after the intra-domain correlation measurement will cause the inter-domain time series aggregation offset, and it is expected to further improve the full-time domain aggregation regression comprehensibility of the significant waveform feature vector of the full-frequency domain ECG signal, thereby improving the accuracy of the measurement detection result obtained by the measurement detection result generator based on the classifier as input.

[0028] Based on this, in the preferred example, inputting the significant waveform feature vector of the full-frequency domain electrocardiogram signal into a classifier-based measurement detection result generator to obtain a measurement detection result specifically includes: Subtract the zero norm of the significant waveform feature vector of the full-frequency domain electrocardiogram signal from the length of the significant waveform feature vector of the full-frequency domain electrocardiogram signal to obtain an isolated zero value of the significant waveform of the full-frequency domain electrocardiogram signal; Calculate the logarithm of the square of the significant isolated zero value of the full-frequency domain electrocardiogram signal waveform and the sum of the significant isolated zero value of the full-frequency domain electrocardiogram signal waveform to obtain the significant waveform superposition information value of the full-frequency domain electrocardiogram signal; Calculate a power function with each eigenvalue of the full-frequency domain electrocardiogram signal significant waveform eigenvector as the base, the difference value of the full-frequency domain electrocardiogram signal significant waveform isolated zero value minus one as the exponent, and perform a point multiplication with the full-frequency domain electrocardiogram signal significant waveform superposition information value to obtain a first full-frequency domain electrocardiogram signal significant waveform intermediate vector; After performing a dot multiplication on the full-frequency domain electrocardiogram signal significant waveform feature vector and the difference between the full-frequency domain electrocardiogram signal significant waveform isolated zero value minus one, the dot multiplication is then performed on the inverse of the full-frequency domain electrocardiogram signal significant waveform isolated zero value to obtain a second full-frequency domain electrocardiogram signal significant waveform intermediate vector; Calculate an exponential function with a natural constant as the base, and each eigenvalue of the second full-frequency domain electrocardiogram signal significant waveform intermediate vector as the exponent to obtain a third full-frequency domain electrocardiogram signal significant waveform intermediate vector; Performing dot addition of the first full-frequency domain electrocardiogram signal significant waveform intermediate vector and the third full-frequency domain electrocardiogram signal significant waveform intermediate vector to obtain an optimized full-frequency domain electrocardiogram signal significant waveform feature vector; and The optimized full-frequency domain electrocardiogram signal significant waveform feature vector is input into a classifier-based measurement detection result generator to obtain a measurement detection result.

[0029] That is, the significant waveform feature vector of the full-frequency domain electrocardiogram signal is, for example, recorded as The optimization process is expressed as: in, is the significant waveform feature vector of the full-frequency domain electrocardiogram signal, It is the isolated zero value of the significant waveform of the full-frequency electrocardiogram signal. It represents the base value of each position characteristic value in the significant waveform characteristic vector of the full-frequency domain electrocardiogram signal The power, is the length of the significant waveform feature vector of the full-frequency domain electrocardiogram signal, is the zero norm of the significant waveform feature vector of the full-frequency domain electrocardiogram signal, It means point multiplication by position. It means adding by position. represents the logarithm operation with base 2, () indicates exponential operation, is the optimized full-frequency domain electrocardiogram signal significant waveform feature vector.

[0030] In this way, the high-dimensional feature manifold of the significant waveform feature vector of the full-frequency domain electrocardiogram signal is represented by the eigenvalue of the feature set as the vector field of the aggregation dimension, and the superposition value of the vector field of the significant waveform feature vector of the full-frequency domain electrocardiogram signal at the isolated zero position is used as the order information to fix the local position of the eigenvalue of its feature set, and a bias for the reversibility of the feature regression distribution field of the significant waveform feature vector of the full-frequency domain electrocardiogram signal is added as a reward to achieve the mapping target tracking of the regression distribution of the significant waveform feature vector of the full-frequency domain electrocardiogram signal for the eigenvalue position, so that the feature set of the significant waveform feature vector of the full-frequency domain electrocardiogram signal can perceive the mapping migration to the aggregate distribution, thereby improving the accuracy of the metrological detection result obtained by the metrological detection result generator based on the classifier of the significant waveform feature vector of the full-frequency domain electrocardiogram signal by improving the full-time domain aggregate regression comprehensibility of the significant waveform feature vector of the full-frequency domain electrocardiogram signal. This will help to more accurately detect whether the patient's heart rate is abnormal, thereby achieving more intelligent medical measurement testing. In this way, human intervention can be effectively reduced and the problems of misdiagnosis, missed diagnosis and inefficiency caused by traditional methods can be avoided.

[0031] As described above, the medical metrology detection device 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a medical metrology detection algorithm. In a possible implementation, the medical metrology detection device 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the medical metrology detection device 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the medical metrology detection device 300 can also be one of the many hardware modules of the wireless terminal.

[0032] Alternatively, in another example, the medical metrology detection device 300 and the wireless terminal may also be separate devices, and the medical metrology detection device 300 may be connected to the wireless terminal via a wired and / or wireless network, and transmit interactive information in a predetermined data format.

[0033] Furthermore, a medical metrology detection method is also provided.

[0034] Figure 3 FIG. 1 is a flow chart of a medical metrology detection method according to an embodiment of the present application. Figure 3 As shown, according to the embodiment of the present application, the medical metrology detection method includes the following steps: S1, acquiring an electrocardiogram signal; S2, performing CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; S3, respectively performing feature extraction on each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; S4, inputting the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; S5, performing aggregation processing on the set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors to obtain significant waveform semantic features of full-frequency domain electrocardiogram signals; S6, determining a metrology detection result based on the significant waveform semantic features of the full-frequency domain electrocardiogram signals, and the metrology detection result is used to indicate whether there is arrhythmia.

[0035] In summary, the medical metrology detection method according to the embodiment of the present application is explained, which collects the electrocardiogram signal of the patient being tested, and introduces a signal processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the electrocardiogram signal, so as to decompose the electrocardiogram signal using the CEEMDAN decomposition technology to extract its intrinsic mode function (IMF) component signal, and then use artificial intelligence technology to learn and characterize the full frequency domain association and aggregation feature representation between the waveform semantics of the frequency band IMF component signal of the electrocardiogram signal, so as to detect whether the heart rate of the patient being tested is abnormal. In this way, artificial intelligence technology can be used to automatically analyze the electrocardiogram signal, thereby realizing more intelligent medical metrology detection. In this way, human intervention can be effectively reduced, and the problems of misdiagnosis, missed diagnosis and low efficiency caused by traditional methods can be avoided.

[0036] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A medical measurement and detection device, characterized in that: include: An electrocardiogram signal acquisition module, used for acquiring electrocardiogram signals; an electrocardiogram signal decomposition module, configured to perform CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; an electrocardiogram component signal waveform feature extraction module, configured to perform feature extraction on each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; an electrocardiogram component signal waveform feature association optimization module, configured to input the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; A full-frequency domain electrocardiogram signal waveform semantic aggregation representation module is used to aggregate the set of the optimized electrocardiogram frequency band IMF component signal waveform feature vectors to obtain the full-frequency domain electrocardiogram signal waveform significant semantic features; The arrhythmia detection module is used to determine a measurement detection result based on the significant waveform semantic features of the full-frequency domain electrocardiogram signal, and the measurement detection result is used to indicate whether arrhythmia exists.

2. The medical measurement and detection equipment according to claim 1, characterized in that: The electrocardiogram component signal waveform feature extraction module is used to: input each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals into a component signal feature extractor based on a one-dimensional convolutional neural network model to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors.

3. The medical measurement and detection equipment according to claim 2, characterized in that: The electrocardiogram component signal waveform feature association optimization module includes: a semantic association score calculation unit, configured to calculate a semantic association score vector between any two electrocardiogram frequency band IMF component signal waveform feature vectors in the set of electrocardiogram frequency band IMF component signal waveform feature vectors to obtain a set of electrocardiogram frequency band IMF component signal waveform semantic association score vectors; A correlation global representation unit, configured to determine an endogenous correlation global representation vector of an electrocardiogram frequency band IMF component signal waveform sequence based on a set of semantic association score vectors of the electrocardiogram frequency band IMF component signal waveforms; an association optimization factor calculation unit, configured to calculate, based on a global representation vector of endogenous correlation of the electrocardiogram frequency band IMF component signal waveform sequence, an association optimization factor of each electrocardiogram frequency band IMF component signal waveform feature vector in a set of electrocardiogram frequency band IMF component signal waveform feature vectors to obtain a sequence of semantic association optimization factors of the electrocardiogram frequency band IMF component signal waveform; The activation unit is used to input the sequence of semantic association optimization factors of the electrocardiogram frequency band IMF component signal waveform Activation function to obtain a sequence of weight factors for semantic association optimization of ECG frequency band IMF component signal waveforms; The electrocardiogram component signal waveform feature optimization unit is used to use each electrocardiogram frequency band IMF component signal waveform semantic association optimization weight factor in the sequence of the electrocardiogram frequency band IMF component signal waveform semantic association optimization weight factors as a weight, and weight each electrocardiogram frequency band IMF component signal waveform feature vector in the set of the electrocardiogram frequency band IMF component signal waveform feature vectors to obtain the set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors.

4. The medical measurement and detection equipment according to claim 3, characterized in that: The semantic association score calculation unit is used to: Cascading any two of the electrocardiogram frequency band IMF component signal waveform feature vectors in the set of electrocardiogram frequency band IMF component signal waveform feature vectors to obtain a set of electrocardiogram frequency band IMF component signal waveform semantic cascade feature vectors; Each electrocardiogram frequency band IMF component signal waveform semantic cascade feature vector in the set of the electrocardiogram frequency band IMF component signal waveform semantic cascade feature vectors is multiplied by the corresponding electrocardiogram frequency band IMF component weight matrix, and then added to the electrocardiogram frequency band IMF component bias vector by position to obtain the set of the electrocardiogram frequency band IMF component signal waveform semantic association score vectors.

5. The medical measurement and detection equipment according to claim 4, characterized in that: The correlation global representation unit is used to calculate the mean vector of the set of semantic association score vectors of the electrocardiogram frequency band IMF component signal waveform as the global representation vector of the endogenous correlation of the electrocardiogram frequency band IMF component signal waveform sequence.

6. The medical measurement and detection equipment according to claim 5, characterized in that: The associated optimization factor calculation unit is used to: Multiplying the global representation vector of the endogenous correlation of the electrocardiogram frequency band IMF component signal waveform sequence by the global weight matrix of the electrocardiogram frequency band IMF component signal waveform to obtain a global representation weight vector of the electrocardiogram frequency band IMF component signal waveform; Multiplying each electrocardiogram frequency band IMF component signal waveform feature vector in the set of electrocardiogram frequency band IMF component signal waveform feature vectors by the corresponding electrocardiogram frequency band IMF component signal waveform local weight matrix to obtain a set of electrocardiogram frequency band IMF component signal waveform local representation weight vectors; Calculating each electrocardiogram frequency band IMF component signal waveform local representation weight vector in the set of electrocardiogram frequency band IMF component signal waveform local representation weight vectors, adding the electrocardiogram frequency band IMF component signal waveform global representation weight vector and the corresponding electrocardiogram frequency band IMF component signal waveform bias vector according to position to obtain a set of electrocardiogram frequency band IMF component signal waveform semantic association vectors; The set of semantic association vectors of the electrocardiogram frequency band IMF component signal waveform is input into a Sigmoid function to obtain a sequence of semantic association optimization factors of the electrocardiogram frequency band IMF component signal waveform.

7. The medical measurement and detection equipment according to claim 6, characterized in that: The full-frequency domain electrocardiogram signal waveform semantic aggregation representation module is used to: cascade the set of the optimized electrocardiogram frequency band IMF component signal waveform feature vectors to obtain the full-frequency domain electrocardiogram signal significant waveform feature vector as the full-frequency domain electrocardiogram signal significant waveform semantic feature.

8. The medical measurement and detection equipment according to claim 7, characterized in that: The arrhythmia detection module is used to: input the significant waveform feature vector of the full-frequency domain electrocardiogram signal into a classifier-based measurement detection result generator to obtain a measurement detection result, and the measurement detection result is used to indicate whether there is an arrhythmia.

9. A medical metrology detection method, characterized in that: include: Acquire electrocardiogram signals; Performing CEEMDAN decomposition on the electrocardiogram signal to obtain a set of electrocardiogram IMF component signals; Performing feature extraction on each electrocardiogram IMF component signal in the set of electrocardiogram IMF component signals to obtain a set of electrocardiogram frequency band IMF component signal waveform feature vectors; Inputting the set of electrocardiogram frequency band IMF component signal waveform feature vectors into a feature association optimization network based on sequence endogenous correlation semantic measurement to obtain a set of optimized electrocardiogram frequency band IMF component signal waveform feature vectors; Aggregating the set of waveform feature vectors of the optimized electrocardiogram frequency band IMF component signal to obtain significant waveform semantic features of the full-frequency domain electrocardiogram signal; Based on the significant waveform semantic features of the full-frequency domain electrocardiogram signal, a measurement detection result is determined, and the measurement detection result is used to indicate whether there is an arrhythmia.

Images