Anesthesia depth classification method and system, storage medium and computer

Abstract

The invention provides an anesthesia depth classification method and system, a storage medium and a computer, and the method comprises the steps: extracting corresponding electroencephalogram data from operation records of a plurality of anesthesia operations based on a preset extraction rule, and carrying out the data preprocessing of the electroencephalogram data through a filtering algorithm, and obtaining the preprocessing data; inputting the preprocessed data into a neural network model to obtain anesthesia concentration data corresponding to the preprocessed data; determining an anesthesia target value corresponding to the preprocessed data according to the operation record, and performing fuzzy processing on the anesthesia target value and the anesthesia concentration data by using a fuzzy processing algorithm to obtain an anesthesia adjustment parameter; and performing model optimization on the neural network model by using the anesthesia adjustment parameters to construct an anesthesia classification model, and performing anesthesia classification on the preprocessed data by using the anesthesia classification model to obtain an anesthesia depth classification result. According to the method, the neural network model is optimized by using the anesthesia adjustment parameters, so that the classification result related to the operation record and the anesthesia depth is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to an anesthesia depth classification method, system, storage medium and computer. Background Art

[0002] With the development and improvement of modern clinical surgery techniques, in clinical surgery, anesthesia is to make the patient lose sensation temporarily as a whole or locally through drugs or other methods to achieve the purpose of painlessness. The effect of anesthesia will have a direct impact on the patient's physical health and also pose a great challenge to anesthesiologists.

[0003] During general anesthesia, anesthesiologists need to estimate the dosage of anesthetic drugs and also plan the injection speed of anesthetic drugs. At present, the judgment process of anesthesia depth is usually achieved through signals such as blood pressure, heart rate, pulse, electroencephalogram, etc. Anesthesiologists make judgments based on the above signals and their own professional knowledge to achieve the precise injection of anesthetic drugs during the anesthesia process. However, this method is relatively dependent on the professional level and clinical experience of anesthesiologists. Therefore, how to achieve the rapid judgment and classification of anesthesia depth is an inevitable problem for anesthesiologists. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide an anesthesia depth classification method, system, storage medium and computer to at least solve the deficiencies in the above technologies.

[0005] The present invention proposes an anesthesia depth classification method, including: Obtain the operation records of several anesthesia surgeries, extract the corresponding electroencephalogram data from the operation records based on a preset extraction rule, and perform data preprocessing on the electroencephalogram data using a filtering algorithm to obtain corresponding preprocessed data; Construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain the anesthesia concentration data corresponding to the preprocessed data; Determine the anesthesia target value corresponding to the preprocessed data according to the operation records, and perform fuzzy processing on the anesthesia target value and the anesthesia concentration data using a fuzzy processing algorithm to obtain corresponding anesthesia adjustment parameters; Use the anesthesia adjustment parameters to optimize the neural network model to construct a corresponding anesthesia classification model, and use the anesthesia classification model to classify the preprocessed data to obtain the anesthesia depth classification results corresponding to each anesthesia surgery.

[0006] Further, the step of extracting the corresponding electroencephalogram data from the operation records based on a preset extraction rule includes: Obtain the amplitude index and the change rate index for extracting EEG data, and construct corresponding extraction rules based on the amplitude index and the change rate index; Extract data from the surgical record according to the extraction rules to obtain corresponding EEG data.

[0007] Further, the neural network model with a multi-level structure at least includes a first convolutional layer, a feature fusion layer, a second convolutional layer, and an adaptive average pooling layer. The step of inputting the preprocessed data into the neural network model to obtain the anesthesia concentration data corresponding to the preprocessed data includes: Input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; Input the fused feature data into the second convolutional layer to calculate the corresponding attention vector, merge the attention vector in the time dimension, and input the obtained merged vector into the adaptive average pooling layer for pooling processing to obtain the anesthesia concentration data corresponding to the preprocessed data.

[0008] Further, the calculation formula for the fused feature data is: ; ; ; In the formula, represents the feature vector obtained by feature extraction of the preprocessed data through the first convolutional layer, represents the feature dimension, represents the time dimension, is the first convolutional sub-layer with a convolutional kernel size of 7 in the first convolutional layer, is the second convolutional sub-layer with a convolutional kernel size of 25 in the first convolutional layer, is the output of the first convolutional sub-layer, is the output of the second convolutional sub-layer, represents the fused feature data, represents the feature fusion layer, which compresses the time dimension to 1; The calculation formula for the merged vector obtained by merging the attention vector in the time dimension is: ; In the formula, represents merging the attention vector in the time dimension, represents the softmax activation function.

[0009] Further, the steps of performing fuzzy processing on the anesthesia target value and the anesthesia concentration data by using a fuzzy processing algorithm to obtain corresponding anesthesia adjustment parameters include: Dividing the anesthesia target value and the anesthesia concentration data into several fuzzy sets, and using a Gaussian function as a membership function to respectively calculate the membership degrees of the anesthesia target value and the anesthesia concentration data belonging to each of the fuzzy sets; Constructing corresponding fuzzy rules according to the membership degrees, and performing defuzzification calculation by using the fuzzy rules, and processing the defuzzification calculation result through an incremental control algorithm to obtain anesthesia adjustment parameters.

[0010] The present invention also proposes an anesthesia depth classification system, including: A data preprocessing module, configured to obtain surgical records of several anesthesia surgeries, extract corresponding electroencephalogram data from the surgical records based on a preset extraction rule, and perform data preprocessing on the electroencephalogram data by using a filtering algorithm to obtain corresponding preprocessed data; A model construction module, configured to construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain anesthesia concentration data corresponding to the preprocessed data; A fuzzy processing module, configured to determine an anesthesia target value corresponding to the preprocessed data according to the surgical records, and perform fuzzy processing on the anesthesia target value and the anesthesia concentration data by using a fuzzy processing algorithm to obtain corresponding anesthesia adjustment parameters; A model optimization module, configured to optimize the neural network model by using the anesthesia adjustment parameters to construct a corresponding anesthesia classification model, and perform anesthesia classification on the preprocessed data by using the anesthesia classification model to obtain anesthesia depth classification results corresponding to each of the anesthesia surgeries.

[0011] Further, the data preprocessing module includes: A rule construction unit, configured to obtain amplitude indexes and change rate indexes for extracting electroencephalogram data, and construct corresponding extraction rules based on the amplitude indexes and the change rate indexes; A data extraction unit, configured to extract data from the surgical records according to the extraction rules to obtain corresponding electroencephalogram data.

[0012] Further, the model construction module includes: A feature extraction unit, configured to input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; A data processing unit for inputting the fused feature data into the second convolutional layer to calculate corresponding attention vectors, merging the attention vectors in the time dimension, and inputting the obtained merged vector into the adaptive average pooling layer for pooling processing to obtain the anesthetic concentration data corresponding to the preprocessed data.

[0013] Further, the fuzzy processing module includes: A fuzzy processing unit for dividing the anesthetic target value and the anesthetic concentration data into several fuzzy sets, and using the Gaussian function as the membership function to calculate the membership degrees of the anesthetic target value and the anesthetic concentration data belonging to each fuzzy set respectively; A parameter calculation unit for constructing corresponding fuzzy rules according to the membership degrees, performing defuzzification calculation using the fuzzy rules, and processing the defuzzification calculation result through an incremental control algorithm to obtain anesthetic adjustment parameters.

[0014] The present invention also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned anesthetic depth classification method is implemented.

[0015] The present invention also provides a computer, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the above-mentioned anesthetic depth classification method is implemented.

[0016] In the anesthetic depth classification method, system, storage medium, and computer of the present invention, by extracting electroencephalogram signals from surgical records and filtering the extracted electroencephalogram signals to avoid the influence of noise in the electroencephalogram signals on the anesthetic depth, the preprocessed data is processed by a constructed neural network model to obtain corresponding anesthetic concentration data, and fuzzy processing is performed by combining the anesthetic concentration data and the anesthetic target value determined by the surgical record, so as to obtain anesthetic adjustment parameters for adjusting the anesthetic depth. Through the anesthetic adjustment parameters, the anesthetic depth is inferred and evaluated to obtain an ideal control effect, and the neural network model is optimized by using the anesthetic adjustment parameters to obtain a classification result related to the surgical record and the anesthetic depth. Description of the Drawings

[0017] Figure 1 It is a flowchart of the anesthetic depth classification method in the first embodiment of the present invention; Figure 2 It is a structural block diagram of the anesthetic depth classification system in the second embodiment of the present invention; Figure 3 It is a structural block diagram of the computer in the third embodiment of the present invention.

[0018] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. Specific Embodiments

[0019] For ease of understanding the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the description of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0021] Embodiment 1 Please refer to Figure 1 , which shows the anesthesia depth classification method in the first embodiment of the present invention. The method specifically includes steps S101 to S105: S101, obtain the surgical records of several anesthesia surgeries, extract the corresponding electroencephalogram data from the surgical records based on a preset extraction rule, and perform data preprocessing on the electroencephalogram data using a filtering algorithm to obtain the corresponding preprocessed data; Further, the step S101 specifically includes steps S1011 to S1012: S1011, obtain the amplitude index and change rate index for extracting electroencephalogram data, and construct the corresponding extraction rule based on the amplitude index and the change rate index; S1012, extract data from the surgical records according to the extraction rule to obtain the corresponding electroencephalogram data.

[0022] In specific implementation, surgical records of several anesthesia surgeries are obtained. Since the electroencephalogram (EEG) signals during anesthesia have higher amplitudes and lower frequencies compared to the awake state, and since EEG signals are subject to noise interference, in this embodiment, recording is started before the surgery and stopped after the surgery. A sliding window with a preset length of 2 s and an overlap of 1 s is set to segment the complete EEG signals in the surgical records, and a noise discrimination algorithm is used to discriminate the segmented signals. The noise discrimination algorithm constructs amplitude indicators and rate-of-change indicators for extracting EEG data. In this embodiment, the threshold of the amplitude indicator is set to 200, and the threshold of the rate-of-change indicator is set to 50. When the amplitude is too large or the rate of change is too drastic, the segment of the signal is judged as noise. All the segmented signals are processed in turn to extract the EEG data related to anesthesia.

[0023] Specifically, a filtering algorithm is used to filter the extracted EEG data to obtain corresponding preprocessed data.

[0024] S102, construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain the anesthesia concentration data corresponding to the preprocessed data; Further, the neural network model with a multi-level structure at least includes a first convolutional layer, a feature fusion layer, a second convolutional layer, and an adaptive average pooling layer. The step S102 specifically includes steps S1021 to S1022: S1021, input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; S1022, input the fused feature data into the second convolutional layer to calculate corresponding attention vectors, merge the attention vectors in the time dimension, and input the obtained merged vectors into the adaptive average pooling layer for pooling processing to obtain the anesthesia concentration data corresponding to the preprocessed data.

[0025] In specific implementation, two convolutional layers are constructed. Each convolutional layer contains two convolutional sub-layers. Among them, the first convolutional sub-layer is used to extract features from EEG signals, and the second convolutional sub-layer is used to expand and compress features. The kernel size of the first convolutional sub-layer is 7, and the kernel size of the second convolutional sub-layer is 25. The stride of each convolutional sub-layer is 2, which is used to downsample the signal. A feature fusion layer is used between the convolutional layers to further perform feature fusion on the signal to improve the accuracy of data processing.

[0026] Specifically, the above-obtained preprocessed data is input into the first convolutional layer for feature extraction, and the extracted feature data is fused through the feature fusion layer to obtain the corresponding fused feature data. The calculation formula for the fused feature data is as follows: ; ; ; In the formula, represents the feature vector obtained by feature extraction of the preprocessed data through the first convolutional layer, represents the feature dimension, represents the time dimension, is the first convolutional sublayer with a convolutional kernel size of 7 in the first convolutional layer, is the second convolutional sublayer with a convolutional kernel size of 25 in the first convolutional layer, is the output of the first convolutional sublayer, is the output of the second convolutional sublayer, represents the fused feature data, represents the feature fusion layer, which compresses the time dimension to 1; Furthermore, the obtained fused feature data is input into the second convolutional layer, and two convolutional sublayers of the second convolutional layer are used to calculate the attention vectors containing small-scale features and large-scale features in the fused feature data respectively. Among them, the small-scale features are the high-frequency, short-time, and local instantaneous activities in the electroencephalogram signal, such as spindle waves, and the large-scale features are the low-frequency, long-time, and global rhythms or network activities in the electroencephalogram signal, such as band energy. The two obtained attention vectors are merged in the time dimension to obtain the corresponding merged vector: ; In the formula, represents merging the attention vectors in the time dimension, represents the softmax activation function.

[0027] Furthermore, the merged vector is weighted into the two convolutional sublayers of the first convolutional layer to obtain the corresponding fused output: ; In the formula, and are the attention vectors calculated by the first convolutional sublayer and the second convolutional sublayer in the first convolutional layer, is the output vector of the adaptive average pooling layer. Based on the same steps, the two convolutional sublayers of the second convolutional layer are calculated to obtain the second output vector ; Perform a second convolution operation on the fused output. This convolution also uses two convolutional layers, both with a kernel size of 7: ; ; In the formula, and respectively represent the third convolutional layer and the fourth convolutional layer; Input the obtained convolution processing result into an adaptive average pooling layer for processing to obtain the corresponding feature vector: ; Use the third convolutional layer and the fourth convolutional layer to perform convolution calculations on the feature vector to obtain the corresponding output value, that is, the anesthetic concentration data corresponding to the preprocessed data: ; ; In the formula, represents the output obtained by performing convolution calculations on the feature vector through the third convolutional layer and the fourth convolutional layer, and respectively represent the output obtained by performing convolution calculations on the feature vector through the third convolutional layer and the output obtained by performing convolution calculations on the feature vector through the fourth convolutional layer, represents the anesthetic concentration data.

[0028] S103. Determine the anesthetic target value corresponding to the preprocessed data according to the surgical record, and use a fuzzy processing algorithm to perform fuzzy processing on the anesthetic target value and the anesthetic concentration data to obtain the corresponding anesthetic adjustment parameter; Further, the step S103 specifically includes steps S1031 to S1032: S1031. Divide the anesthetic target value and the anesthetic concentration data into several fuzzy sets, and use the Gaussian function as the membership function to calculate the membership degrees of the anesthetic target value and the anesthetic concentration data belonging to each fuzzy set respectively; S1032. Construct corresponding fuzzy rules according to the membership degrees, and use the fuzzy rules to perform defuzzification calculations, and process the defuzzification calculation results through an incremental control algorithm to obtain the anesthetic adjustment parameter.

[0029] In specific implementation, the anesthesia target value corresponding to the preprocessed data is determined according to the surgical record. Here, the anesthesia target value is the theoretical anesthesia value corresponding to this surgical record. When the patient of this surgical record reaches this anesthesia target value, the corresponding anesthesia depth is theoretically the optimal anesthesia depth that meets the surgical requirements. Calculate the deviation value and the corresponding deviation change rate between the obtained anesthesia concentration data and the anesthesia target value, and use the deviation value and the deviation change rate as inputs and input them into the preset input layer, so that each node of this layer is directly connected to each component of the input data. Each node of this layer The input is expressed as: ; Divide the obtained input above into 7 fuzzy sets, and use the Gaussian function as the membership function to calculate the membership degrees of the anesthesia target value and the anesthesia concentration data belonging to each fuzzy set respectively: ; In the formula, , , and respectively represent the mean and standard deviation of the membership function of the th input variable belonging to the th fuzzy set; Furthermore, corresponding fuzzy rules are constructed according to the membership degrees. The output of each node is calculated as the product of all input signals at this node: ; In the formula, represents the number of fuzzy rules, ; Specifically, use the obtained fuzzy rules for defuzzification calculation, process the defuzzification calculation result through an incremental control algorithm (in this embodiment, this incremental control algorithm uses the PID control algorithm) to obtain the corresponding control law, and use a supervised learning algorithm (in this embodiment, this supervised learning algorithm includes the Delta algorithm, the LSM algorithm, etc.) to correct the obtained control rules to obtain the corresponding output weights, that is, the parameters for anesthesia adjustment.

[0030] S104, use the anesthesia adjustment parameters to optimize the neural network model to construct the corresponding anesthesia classification model, and use the anesthesia classification model to classify the preprocessed data to obtain the anesthesia depth classification results corresponding to each anesthesia surgery.

[0031] In specific implementation, a preset anesthesia classification index is constructed, where the anesthesia classification index includes the anesthesia depth corresponding to the calculated anesthesia value. The anesthesia classification index and anesthesia adjustment parameters are input into a neural network model for model optimization, so as to obtain a corresponding anesthesia classification model. The anesthesia classification model is used to classify the preprocessed data, so as to obtain the anesthesia depth classification results corresponding to each anesthesia surgery. In summary, for the anesthesia depth classification method in the above embodiments of the present invention, by extracting electroencephalogram signals from surgical records and filtering the extracted electroencephalogram signals to avoid the influence of noise in the electroencephalogram signals on the anesthesia depth, the preprocessed data is processed by the constructed neural network model to obtain corresponding anesthesia concentration data, and fuzzy processing is performed in combination with the anesthesia concentration data and the anesthesia target value determined by the surgical records, so as to obtain anesthesia adjustment parameters for adjusting the anesthesia depth. The anesthesia depth is inferred and evaluated through the anesthesia adjustment parameters, so as to obtain an ideal control effect. The model of the neural network model is optimized by using the anesthesia adjustment parameters to obtain a classification result related to the surgical record and the anesthesia depth.

[0032] Embodiment 2 On the other hand, the present invention also proposes an anesthesia depth classification system. Please refer to Figure 2 , which shows the anesthesia depth classification system in the second embodiment of the present invention. The system includes: A data preprocessing module 11, configured to obtain surgical records of several anesthesia surgeries, extract corresponding electroencephalogram data from the surgical records based on a preset extraction rule, and perform data preprocessing on the electroencephalogram data by using a filtering algorithm to obtain corresponding preprocessed data; Further, the data preprocessing module 11 includes: A rule construction unit, configured to obtain amplitude indexes and change rate indexes for extracting electroencephalogram data, and construct corresponding extraction rules based on the amplitude indexes and the change rate indexes; A data extraction unit, configured to extract data from the surgical records according to the extraction rules to obtain corresponding electroencephalogram data.

[0033] A model construction module 12, configured to construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain anesthesia concentration data corresponding to the preprocessed data; Further, the model construction module 12 includes: A feature extraction unit, configured to input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; A data processing unit, configured to input the fused feature data into the second convolutional layer to calculate corresponding attention vectors, merge the attention vectors in the time dimension, and input the obtained merged vector into the adaptive average pooling layer for pooling processing to obtain the anesthetic concentration data corresponding to the preprocessed data.

[0034] A fuzzy processing module 13, configured to determine the anesthetic target value corresponding to the preprocessed data according to the surgical record, and perform fuzzy processing on the anesthetic target value and the anesthetic concentration data by using a fuzzy processing algorithm to obtain corresponding anesthetic adjustment parameters; Further, the fuzzy processing module 13 includes: A fuzzy processing unit, configured to divide the anesthetic target value and the anesthetic concentration data into several fuzzy sets, and use a Gaussian function as a membership function to calculate the membership degrees of the anesthetic target value and the anesthetic concentration data belonging to each of the fuzzy sets respectively; A parameter calculation unit, configured to construct corresponding fuzzy rules according to the membership degrees, perform defuzzification calculation by using the fuzzy rules, and process the defuzzification calculation result through an incremental control algorithm to obtain anesthetic adjustment parameters.

[0035] A model optimization module 14, configured to optimize the neural network model by using the anesthetic adjustment parameters to construct a corresponding anesthetic classification model, and classify the preprocessed data by using the anesthetic classification model to obtain anesthetic depth classification results corresponding to each of the anesthetic surgeries.

[0036] The functions or operation steps implemented when the above-mentioned modules and units are executed are substantially the same as those in the above method embodiments, and will not be described in detail herein.

[0037] The anesthetic depth classification system provided by the embodiments of the present invention has the same implementation principle and technical effects as those in the foregoing method embodiments. For a brief description, for the parts not mentioned in the system embodiments, reference may be made to the corresponding content in the foregoing method embodiments.

[0038] Embodiment III The present invention further provides a computer. Please refer to Figure 3 , which shows the computer in the third embodiment of the present invention, including a memory 10, a processor 20, and a computer program 30 stored on the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, the above-mentioned anesthetic depth classification method is implemented.

[0039] Among them, the memory 10 includes at least one type of storage medium, and the storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, magnetic disk, optical disc, etc. The memory 10 can be an internal storage unit of a computer in some embodiments, such as the hard disk of the computer. The memory 10 can also be an external storage device in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. Further, the memory 10 can also include both an internal storage unit of a computer and an external storage device. The memory 10 can be used not only to store application software installed in the computer and various types of data, but also to temporarily store data that has been output or will be output.

[0040] Among them, the processor 20 can be an Electronic Control Unit (ECU, also known as a vehicle computer), a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments, and is used to run the program code stored in the memory 10 or process data, such as executing an access restriction program, etc.

[0041] It should be noted that Figure 3 the structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown in the figure, or combine certain components, or have a different component layout.

[0042] An embodiment of the present invention also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the anesthesia depth classification method as described above.

[0043] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus, or device.

[0044] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then storing it in a computer memory.

[0045] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0046] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0047] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for classifying anesthesia depth, characterized in that, Including: Obtain the operation records of a number of anesthetic operations, extract the corresponding electroencephalogram data from the operation records based on a preset extraction rule, and use a filtering algorithm to perform data preprocessing on the electroencephalogram data to obtain corresponding preprocessed data; Construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain the anesthetic concentration data corresponding to the preprocessed data; Determine the anesthetic target value corresponding to the preprocessed data according to the operation record, and use a fuzzy processing algorithm to perform fuzzy processing on the anesthetic target value and the anesthetic concentration data to obtain a corresponding anesthetic adjustment parameter; Use the anesthetic adjustment parameter to optimize the neural network model to construct a corresponding anesthetic classification model, and use the anesthetic classification model to perform anesthetic classification on the preprocessed data to obtain the anesthetic depth classification results corresponding to each anesthetic operation.

2. The anesthesia depth classification method according to claim 1, characterized in that The step of extracting the corresponding electroencephalogram data from the operation record based on a preset extraction rule includes: Obtain the amplitude index and change rate index for extracting electroencephalogram data, and construct a corresponding extraction rule based on the amplitude index and the change rate index; Extract data from the operation record according to the extraction rule to obtain the corresponding electroencephalogram data.

3. The anesthesia depth classification method according to claim 1, wherein The neural network model with a multi-level structure at least includes a first convolutional layer, a feature fusion layer, a second convolutional layer, and an adaptive average pooling layer. The step of inputting the preprocessed data into the neural network model to obtain the anesthetic concentration data corresponding to the preprocessed data includes: Input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; Input the fused feature data into the second convolutional layer to calculate a corresponding attention vector, merge the attention vector in the time dimension, and input the obtained merged vector into the adaptive average pooling layer for pooling processing to obtain the anesthetic concentration data corresponding to the preprocessed data.

4. The method for classifying the depth of anesthesia according to claim 3, characterized in that, The calculation formula for the fused feature data is: ； ； ； In the formula, represents the feature vector obtained by the feature extraction of the preprocessed data through the first convolutional layer, represents the feature dimension, represents the time dimension, is the first convolutional sublayer with a convolutional kernel size of 7 in the first convolutional layer, is the second convolutional sublayer with a convolutional kernel size of 25 in the first convolutional layer, is the output of the first convolutional sublayer, is the output of the second convolutional sublayer, represents the fused feature data, represents the feature fusion layer, which compresses the time dimension to 1; The calculation formula for the merged vector obtained by merging the attention vector in the time dimension is: ； In the formula, represents the merging of the attention vector in the time dimension, represents the softmax activation function.

5. The method for classifying anesthesia depth according to claim 1, characterized in that The step of using a fuzzy processing algorithm to perform fuzzy processing on the anesthetic target value and the anesthetic concentration data to obtain a corresponding anesthetic adjustment parameter includes: Divide the anesthetic target value and the anesthetic concentration data into a number of fuzzy sets, and use a Gaussian function as the membership function to calculate the membership degrees of the anesthetic target value and the anesthetic concentration data belonging to each fuzzy set respectively; Construct a corresponding fuzzy rule according to the membership degrees, and perform defuzzification calculation using the fuzzy rule, and process the defuzzification calculation result through an incremental control algorithm to obtain the anesthetic adjustment parameter.

6. An anesthesia depth classification system, characterized in that, Including: A data preprocessing module, configured to obtain surgical records of a number of anesthesia surgeries, extract corresponding electroencephalogram data from the surgical records based on a preset extraction rule, and perform data preprocessing on the electroencephalogram data by using a filtering algorithm to obtain corresponding preprocessed data; A model construction module, configured to construct a neural network model with a multi-level structure, and input the preprocessed data into the neural network model to obtain anesthesia concentration data corresponding to the preprocessed data; A fuzzy processing module, configured to determine an anesthesia target value corresponding to the preprocessed data according to the surgical record, and perform fuzzy processing on the anesthesia target value and the anesthesia concentration data by using a fuzzy processing algorithm to obtain corresponding anesthesia adjustment parameters; A model optimization module, configured to optimize the neural network model by using the anesthesia adjustment parameters to construct a corresponding anesthesia classification model, and perform anesthesia classification on the preprocessed data by using the anesthesia classification model to obtain anesthesia depth classification results corresponding to each of the anesthesia surgeries.

7. The depth of anesthesia classification system according to claim 6, wherein The data preprocessing module includes: A rule construction unit, configured to obtain amplitude indexes and change rate indexes for extracting electroencephalogram data, and construct a corresponding extraction rule based on the amplitude indexes and the change rate indexes; A data extraction unit, configured to extract data from the surgical record according to the extraction rule to obtain corresponding electroencephalogram data.

8. The anesthesia depth classification system according to claim 6, characterized in that, The model construction module includes: A feature extraction unit, configured to input the preprocessed data into the first convolutional layer for feature extraction, and fuse the extracted feature data through the feature fusion layer to obtain corresponding fused feature data; A data processing unit, configured to input the fused feature data into the second convolutional layer to calculate a corresponding attention vector, merge the attention vector in the time dimension, and input the obtained merged vector into the adaptive average pooling layer for pooling processing to obtain anesthesia concentration data corresponding to the preprocessed data.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the anesthesia depth classification method according to any one of claims 1 to 5.

10. A computer, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the anesthesia depth classification method according to any one of claims 1 to 5.

Images