Physiological parameter dynamic evaluation method

Abstract

The invention discloses a physiological parameter dynamic evaluation method. The method comprises the following steps: acquiring physiological parameter information of a user; the physiological parameter information of the user comprises electroencephalogram information, electrocardio information, near-infrared information and blood information; preprocessing the user physiological parameter information to obtain preprocessed physiological parameter information; the pre-processed physiological parameter information comprises pre-processed electroencephalogram information, pre-processed electrocardio information, pre-processed near-infrared information and pre-processed blood information; training a preset physiological parameter dynamic evaluation model by utilizing the preprocessed physiological parameter information to obtain an optimized physiological parameter dynamic evaluation model; and processing the to-be-processed user physiological parameter information by using the optimized physiological parameter dynamic evaluation model to obtain a physiological parameter dynamic evaluation result. According to the method, only common noninvasive parameters are used, the physiological parameters of the user are dynamically evaluated, sufficient early warning time is reserved for clinical intervention, and sufficient time can be provided for a doctor to design a patient treatment scheme.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method for dynamically evaluating physiological parameters. Background Art

[0002] Physiological parameter information can reflect the physical condition of the human body, and the evaluation of physiological parameter information can realize dynamic early warning of the physical state. When treating patients with liver failure, it is often necessary to keep track of the patient's vital signs at all times to prevent the patient's condition from worsening. The existing vital sign detectors often only detect and process each type of physiological signal separately. For critically ill patients, their physiological parameters often show variable and time-varying characteristics. It is difficult to quickly and accurately monitor the user's vital signs based on only one type of physiological parameter. How to fuse and process multiple types of physiological signals of patients and quickly and accurately monitor their vital signs to improve the quality and efficiency of patient care is an issue that needs to be solved urgently. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method for dynamic evaluation of physiological parameters, which uses artificial intelligence methods to dynamically evaluate the physiological parameters of users, reserve sufficient warning time for clinical intervention, and provide sufficient time for doctors to design patient treatment plans.

[0004] In order to solve the above technical problems, an embodiment of the present invention discloses a method for dynamic early warning of adverse prognosis of sepsis patients, the method comprising:

[0005] S1, obtaining user physiological parameter information; the user physiological parameter information includes EEG information, ECG information, near infrared information and blood information;

[0006] S2, preprocessing the user's physiological parameter information to obtain preprocessed physiological parameter information; the preprocessed physiological parameter information includes preprocessed EEG information, preprocessed ECG information, preprocessed near-infrared information and preprocessed blood information;

[0007] S3, using the pre-processed physiological parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model;

[0008] S4, using the optimized physiological parameter dynamic evaluation model, processing the user physiological parameter information to be processed to obtain a physiological parameter dynamic evaluation result.

[0009] As an optional implementation manner, in an embodiment of the present invention, the preprocessing of the user physiological parameter information to obtain preprocessed physiological parameter information includes:

[0010] S21, performing abnormality check processing on the user's physiological parameter information to obtain first physiological parameter information;

[0011] S22, normalizing the first physiological parameter information to obtain second physiological parameter information;

[0012] S23, performing denoising processing on the second physiological parameter information to obtain pre-processed physiological parameter information.

[0013] As an optional implementation manner, in an embodiment of the present invention, the denoising process is performed on the second physiological parameter information to obtain pre-processed physiological parameter information, including:

[0014] S231, processing the second physiological parameter information to obtain a real component and an imaginary component;

[0015] S232, performing wavelet transform on the real component to obtain first wavelet parameter information;

[0016] S233, performing wavelet transform on the imaginary component to obtain second wavelet parameter information;

[0017] S234, filtering the first wavelet parameter information to obtain third wavelet parameter information;

[0018] S235, filtering the second wavelet parameter information to obtain fourth wavelet parameter information;

[0019] S236, reconstructing the third wavelet parameter information and the fourth wavelet parameter information to obtain preprocessed physiological parameter information.

[0020] As an optional implementation, in an embodiment of the present invention, the use of the preprocessed physiological parameter information to train a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model includes:

[0021] S31, extracting features from the preprocessed physiological parameter information to obtain physiological parameter feature information; the physiological parameter feature information includes EEG feature parameter information, ECG feature parameter information, near infrared feature parameter information and blood feature parameter information;

[0022] S32, fusing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain fused characteristic parameter information;

[0023] S33, using the fused characteristic parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model.

[0024] As an optional implementation manner, in an embodiment of the present invention, the feature extraction of the pre-processed physiological parameter information to obtain physiological parameter feature information includes:

[0025] S311, extracting features from the preprocessed EEG information to obtain EEG feature parameter information;

[0026] S312, performing feature extraction on the preprocessed ECG information to obtain ECG feature parameter information;

[0027] S313, performing feature extraction on the pre-processed near infrared information to obtain near infrared feature parameter information;

[0028] S314, performing feature extraction on the pre-processed blood information to obtain blood feature parameter information.

[0029] As an optional implementation, in the embodiment of the present invention, the feature extraction of the pre-processed EEG information to obtain EEG feature parameter information includes:

[0030] S3111, extracting time-frequency features from the preprocessed EEG information to obtain time-frequency feature parameters;

[0031] The time-frequency characteristic parameter expression is:

[0032]

[0033] In the formula, S(m,n) is the time-frequency characteristic parameter, x(t) represents the preprocessed EEG information, g(i-nT) represents the window function, T represents the sliding window length, N represents the discrete Fourier transform length, m, n represent the time-frequency characteristic parameter frequency and time respectively, L represents the preprocessed EEG information length, and i represents the preprocessed EEG information sample point;

[0034] S3112, filtering the time-frequency characteristic parameters to obtain EEG characteristic parameter information, including:

[0035] Set the initial threshold δ th The target interval is [α, β], where 0≤α<β≤max{S(m,n)}, and S(m,n) is the time-frequency feature parameter;

[0036] Divide the target interval [α, β] into L equal parts;

[0037] The initial threshold δ th Set to α, according to The step size is incremented and iterated to obtain the threshold

[0038] Utilizing Thresholds Calculate the time-frequency characteristic parameters after filtering for each iteration

[0039]

[0040] Among them, S(m,n) represents the time-frequency characteristic parameters, represents the time-frequency characteristic parameters after filtering, a and b represent the time and frequency parameters in the threshold;

[0041] Calculate the difference between the time-frequency feature parameters obtained in each iteration and the previous iteration i1=1,2,…, i1 is the number of iterations;

[0042] The difference Perform curve fitting to obtain curve C. When curve C has an inflection point, the corresponding threshold is the optimal threshold.

[0043] The optimal threshold is used to perform optimal threshold filtering on the time-frequency feature parameter S(m,n) to obtain the filtered time-frequency feature parameter, which is the EEG feature parameter information.

[0044] As an optional implementation, in an embodiment of the present invention, the feature extraction of the preprocessed ECG information to obtain ECG feature parameter information includes:

[0045] S3121, processing the pre-processed ECG information to obtain fuzzy parameter information; the fuzzy parameter information is expressed as:

[0046]

[0047] in, x(t) is the preprocessed ECG information, A x (θ,τ) is the fuzzy parameter information, t is the time variable, τ is the time shift variable, θ is the frequency variable corresponding to τ, and * indicates taking the conjugate;

[0048] S3122, performing Hadamard product calculation on the fuzzy parameter information and a preset kernel function to obtain ECG characteristic parameter information;

[0049] The Hadamard product calculation formula is:

[0050]

[0051] Among them, C x is the ECG characteristic parameter information, the integration range is -∞~∞, g(θ,τ) is the preset kernel function, a=0.001,β=50,u,t is the time variable, τ is the time shift variable, Ω is the frequency variable corresponding to u, and θ is the frequency variable corresponding to τ.

[0052] As an optional implementation manner, in an embodiment of the present invention, the feature extraction of the pre-processed near-infrared information to obtain near-infrared feature parameter information includes:

[0053] S3131, using a near infrared information analysis model, performing time-frequency processing on the near infrared information to obtain first near infrared characteristic parameter information;

[0054] The near infrared information analysis model expression is:

[0055]

[0056] In the formula, t represents time, ω represents frequency, x() represents near-infrared information, τ represents delay, W(t,ω) is the first near-infrared characteristic parameter information, and * represents conjugation;

[0057] S3132, using a preset filtering model, filtering the first near infrared characteristic parameter information W(t,ω) to obtain near infrared characteristic parameter information

[0058] The preset filtering model expression is:

[0059]

[0060] Where W(τ,θ) represents the first near-infrared characteristic parameter information, W1(t-τ,ω-θ) is the two-dimensional template, and are two Gaussian functions, representing the spatial domain kernel and the pixel domain kernel respectively, τ is the time domain variable, and θ is the frequency domain variable;

[0061] and The specific expression is:

[0062]

[0063] In the formula, a and b represent the horizontal and vertical coordinates of the input pixel, i and j represent the coordinates of the center of the two-dimensional template, and σ s , σ r Represents the standard deviation of the Gaussian function.

[0064] As an optional implementation manner, in an embodiment of the present invention, the feature extraction of the pre-processed blood information to obtain blood feature parameter information includes:

[0065] S3141, performing matrix decomposition on the pre-processed blood information to obtain a left singular matrix, a singular value matrix, and a right singular matrix;

[0066] The matrix decomposition expression is:

[0067] Y=UAV

[0068] Among them, U is the left singular matrix, Y is the pre-processed blood information, A is the singular value matrix, V is the right singular matrix, U and V are both orthogonal matrices, and A is a diagonal matrix;

[0069] S3142, performing difference calculation on the left singular matrix and the right singular matrix to obtain a difference matrix;

[0070] S3143, performing cross-correlation calculation on the difference matrix and the singular value matrix to obtain blood characteristic parameter information.

[0071] As an optional implementation, in an embodiment of the present invention, the fusion of the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information and the blood characteristic parameter information to obtain fused characteristic parameter information includes:

[0072] S3211, using a data centralization model, processing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain centralization parameter information;

[0073] The model expression in the data center is:

[0074]

[0075] The centralized parameter information is X=[w1-u, w2-u, w3-u, w4-u], w1 is EEG characteristic parameter information, w2 is ECG characteristic parameter information, w3 is near-infrared characteristic parameter information, w4 is blood characteristic parameter information, a1 is the weighting coefficient of w1, a2 is the weighting coefficient of w2, a3 is the weighting coefficient of w3, a4 is the weighting coefficient of w4, a1+a2+a3+a4=1;

[0076] S3212, processing the centralization parameter information to obtain a covariance matrix;

[0077] The covariance matrix expression is:

[0078] ∑=XX T

[0079] Where Σ is the covariance matrix and T is the transpose;

[0080] S3213, performing eigenvalue decomposition on the covariance matrix to obtain m eigenvalues;

[0081] S3214, sorting the m eigenvalues, and extracting eigenvectors corresponding to the first k eigenvalues;

[0082] S3215, using the feature vector reconstruction model, processing the feature vectors corresponding to the first k eigenvalues ​​to obtain fused feature parameter information;

[0083] The feature vector reconstruction model expression is:

[0084] Y=[u1X,u2X,…,u k X]

[0085] Among them, Y is the fusion feature parameter information, [u1,u2,…,u k ] is the eigenvector corresponding to the first k eigenvalues.

[0086] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0087] The present invention discloses a method for dynamic evaluation of physiological parameters, which includes: obtaining physiological parameter information of a user; mining and analyzing multiple physiological indicators to obtain dynamic evaluation results of physiological parameters. The method of the present invention uses only commonly used non-invasive parameters to dynamically evaluate the physiological parameters of the user, which reserves sufficient warning time for clinical intervention and can provide sufficient time for doctors to design treatment plans for patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0089] Figure 1 It is a flow chart of a method for dynamic evaluation of physiological parameters disclosed in an embodiment of the present invention;

[0090] Figure 2 It is a schematic diagram of the structure of the preset LSTM neural network disclosed in the embodiment of the present invention. DETAILED DESCRIPTION

[0091] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0092] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0093] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0094] The present invention discloses a method and device for dynamic evaluation of physiological parameters, the method comprising: obtaining physiological parameter information of a user; the physiological parameter information of the user comprises EEG information, ECG information, near infrared information and blood information; preprocessing the physiological parameter information of the user to obtain preprocessed physiological parameter information; the preprocessed physiological parameter information comprises preprocessed EEG information, preprocessed ECG information, preprocessed near infrared information and preprocessed blood information; using the preprocessed physiological parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model; using the optimized physiological parameter dynamic evaluation model, processing the physiological parameter information of the user to be processed to obtain a physiological parameter dynamic evaluation result. The method of the present invention uses only commonly used non-invasive parameters to dynamically evaluate the physiological parameters of the user, reserving sufficient warning time for clinical intervention, and providing sufficient time for doctors to design patient treatment plans. The following are detailed descriptions.

[0095] Example

[0096] See also Figure 1 , Figure 1 : is a flow chart of a method for dynamic evaluation of physiological parameters disclosed in an embodiment of the present invention. Figure 1 The described physiological parameter dynamic evaluation method is applied to the field of artificial intelligence technology. The physiological parameter information to be processed is processed to obtain dynamic evaluation results, which can provide sufficient time for doctors to design patient treatment plans. The embodiments of the present invention are not limited thereto. Figure 1 As shown, the physiological parameter dynamic evaluation method may include the following operations:

[0097] S1, obtaining user physiological parameter information; the user physiological parameter information includes EEG information, ECG information, near infrared information and blood information;

[0098] The EEG information is collected by an EEG acquisition unit, which includes a dry electrode sensor and a signal amplifier. The collected EEG information is stored in a computer through a data line;

[0099] ECG information is collected by an ECG instrument and stored in a computer via a data line;

[0100] The near-infrared laser light source and avalanche diode receiver are used to collect near-infrared information. The laser light source is a 730nm and 850nm near-infrared laser light source. The collected near-infrared information is stored in the computer through a data line.

[0101] Blood information includes: platelets, blood sodium, ALT, AST, CHE, TBiL, DBiL, IBiL, PTA, INR, procalcitonin, IL6 and CRP. Blood information is collected through routine blood tests and biochemical analysis and entered into the system manually by nurses or obtained through online connection with the hospital database;

[0102] S2, preprocessing the user's physiological parameter information to obtain preprocessed physiological parameter information; the preprocessed physiological parameter information includes preprocessed EEG information, preprocessed ECG information, preprocessed near-infrared information and preprocessed blood information;

[0103] S3, using the pre-processed physiological parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model;

[0104] The preset physiological parameter dynamic evaluation model is LSTM neural network:

[0105] Optionally, the pre-processed physiological parameter information is used to train a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model using the following method:

[0106] (1) Denoising preprocessing: A portion of the preprocessed physiological parameter information is used to train the CNN denoising model, and the remaining data is preprocessed with the trained CNN denoising model. The denoised data is output as the training set of the LSTM neural network, and the remaining preprocessed physiological parameter information for training the CNN denoising model is used as the validation set.

[0107] The CNN denoising model is a fully convolutional network consisting of 16 convolutional layers. The size of the convolution kernel of the first layer is 9×8, with 18 convolution kernels. The 2nd to 13th convolutional layers are a group of 3 layers repeated 4 times, with convolution kernel widths of 5, 9, and 9 respectively, and the number of convolution kernels is 64. The convolution kernel width of the last convolution layer is 129 and there is only 1 convolution kernel. In this network, convolution is performed only in the frequency dimension, and for all layers except the first layer, the convolution kernel width along the time dimension is set to 1. The convolution layer passes through a batch normalization layer and a ReLU activation function layer, and is finally output through a regression layer.

[0108] (2) Setting the fitness function: Use the training set to train the LSTM neural network model, and the inverse of the accuracy of the LSTM neural network that reaches the maximum number of training times on the validation set is used as the fitness function.

[0109] (3) Initialize the parameters of the quantum sailfish algorithm and the quantum positions of the quantum sailfish and quantum sardines. Each quantum position is a three-dimensional vector, which represents the number of hidden layer neurons, the initial learning rate, and the number of training times of the preset LSTM neural network model.

[0110] (4) Iterative update of quantum positions: Update the quantum positions of quantum swordfish and quantum sardines, calculate the fitness values, determine the optimal fitness values ​​of individuals, and iteratively update the population.

[0111] (5) After the quantum sailfish algorithm is iterated, the optimal position is taken as the hyperparameter of the LSTM neural network model to obtain the dynamic evaluation model of optimized physiological parameters.

[0112] The quantum sailfish algorithm combines the advantages of the quantum evolution theory of sailfish and quantum computing. In the quantum sailfish algorithm, quantum mechanisms design a new way to update the positions of sailfish and sardines, which can improve the global search ability of sailfish.

[0113] The quantum sailfish algorithm evolves through the quantum states of sailfish and sardine schools. There is a sailfish population consisting of N q Each sailfish has its own quantum position and corresponding position. Therefore, when solving a d-dimensional problem, the quantum position of the i-th sailfish is defined as q i =[q i,1 ,q i,2 ,...,q i,d ], where 0<q i,j <1, i = 1, 2, ..., N q , j = 1, 2, ..., d. There is a sardine school with an initial population size of Where p is the ratio of the initial swordfish population to the sardine population, and the quantum position of the kth sardine is defined as s k =[sk,1 ,s k,2 ,...,s k,d ], where 0<s k,j <1, j=1,2,...,d.

[0114] Map it to a position in the solution space, and the mapping relationship is defined as follows:

[0115]

[0116] Among them, x min =[x min,1 ,x min ,2,...,x min,d ] is the lower boundary of the problem, x max =[x max,1 ,x max ,2,...,x max,d ] is the upper boundary of the problem, and “⊙” represents the Hadamard product, which is the multiplication of the elements in the corresponding dimensions of the two vectors.

[0117] Each position of the swordfish and sardines is a feasible solution to the problem, and the quality of the feasible solution can be calculated by the fitness function f(·). The minimum fitness value of the swordfish population in the g-th iteration is the optimal fitness value. Determine its quantum position as the optimal quantum position of the current iteration The optimal fitness value in the entire swordfish population up to the gth iteration is the global optimal fitness value Its location is the elite sailfish quantum location The minimum fitness value of the sardine population in the gth iteration is the optimal fitness value Determine its quantum position as the injured sardine quantum position of the current iteration

[0118] The sailfish population, led by the elite sailfish, attacks the sardine school and approaches the injured sardines to complete the predation. Therefore, the next generation quantum position of each sailfish is affected by the elite sailfish and the injured sardines. The jth quantum bit of the i-th sailfish in the g+1th iteration is updated using a simulated quantum revolving gate.

[0119]

[0120] Among them, r1 is a uniform random number between [0,1], r2 is a Gaussian distribution random number with a mean of 0 and a variance of 1, and λ g+1 A coefficient is defined as follows

[0121] λ g+1 =2×rand×P g+1-P g+1

[0122] in is the density of sardines, indicating that the number of sardines will decrease significantly with the predation of swordfish in each iteration. is the population size of sardines at the g+1th iteration, and rand is a uniform random number between [0,1].

[0123] As the sailfish continue to attack the sardines, the sardines' energy storage will be reduced, reducing their ability to detect the sailfish's location information, which will affect the sardines' escape, and eventually they will be attacked and injured by the sailfish, becoming the sailfish's prey. Therefore, the jth qubit of the kth sardine in the g+1th iteration is updated as follows:

[0124]

[0125] Among them, r3 is a uniform random number between [0,1], and r4 is a Gaussian distribution random number with a mean of 0 and a variance of 1. g+1 is the energy of the sardines at the g+1th iteration, and its expression is

[0126] A g+1 =E×[1-2×(g+1)×ξ]

[0127] Where E is the total energy of sardines and ξ is the energy attenuation coefficient.

[0128] Sardines injured in the final stage of predation will be caught by the swordfish. When the fitness value of the sardine is better than that of the swordfish, it will be caught, and each swordfish will catch at most one sardine in each iteration. In this case, the quantum position of the swordfish is replaced by the quantum position of the captured sardine. but The caught sardines are removed from the sardine population, so the sardine population is Where D g+1 is the number of sardines caught by swordfish.

[0129] The algorithm terminates when the maximum number of iterations is reached or all sardines are caught, and finally outputs the position of the elite swordfish. The algorithm time complexity is O(n 3 ).

[0130] LSTM neural network is a special form of recurrent neural network (RNN), which is designed to solve the long-term dependency problem of RNN. Figure 2As shown in the figure, an LSTM unit contains a memory cell and three gates. The LSTM neural network uses a gating mechanism to control the accumulation rate during information transmission, and can also selectively forget previous information, so that the gradient will not disappear or explode. The gating mechanism in the LSTM neural network unit mainly includes three parts: forget gate, input gate and output gate.

[0131] The calculation process of the LSTM neural network memory unit can be expressed as:

[0132] f t =sigmoid(W f ⊙x t +W f ⊙h t-1 +b f )

[0133] i t =sigmoid(W i ⊙x t +W i ⊙h t-1 +b i )

[0134] o t =sigmoid(W o ⊙x t +W o ⊙h t-1 +b o )

[0135]

[0136] where f t 、i t and t Respectively represent the outputs of the forget gate, input gate, and output gate; x t is the input of the LSTM unit memory cell at time t; h t-1 The output of the previous LSTM unit memory cell; W represents the temporary memory cell state of the LSTM unit at time t; f , W i , W o and W c Indicates that the weight matrix corresponds to the forget gate, input gate, output gate and neuron update matrix, b f 、b i 、b o and b c are the corresponding bias values ​​respectively. According to the above formula, the current memory cell state and output of the LSTM unit can be expressed as:

[0137]

[0138] h t =o t ⊙tanh(c t )

[0139] This is the forward calculation process of the LSTM neural network. This special structure enables it to learn long-term dependencies and has been widely used in fields such as text analysis and time series prediction.

[0140] When training LSTM, the input data of the input layer is first forward propagated through the LSTM unit of the hidden layer, and the actual output of the neural network at the mth iteration is obtained through the output layer as a m , the probability that the final output of the network belongs to a certain category is calculated using the softmax function

[0141]

[0142] Where k∈1,2,...,K, K represents the number of network output categories. According to the true value y=[y1,y2,...,y K ], the loss function used to train the LSTM network is cross entropy:

[0143]

[0144] S4, using the optimized physiological parameter dynamic evaluation model, processing the user physiological parameter information to be processed to obtain a physiological parameter dynamic evaluation result.

[0145] The dynamic evaluation results of physiological parameters include the patient's corresponding mortality rate within the next 7 days, mortality rate within 30 days, expected ICU stay, and expected total hospital stay.

[0146] Optionally, the preprocessing the user physiological parameter information to obtain preprocessed physiological parameter information includes:

[0147] S21, performing abnormality check processing on the user's physiological parameter information to obtain first physiological parameter information;

[0148] Abnormality checking is a conventional data processing method in this field, and this embodiment does not limit it;

[0149] S22, normalizing the first physiological parameter information to obtain second physiological parameter information;

[0150] Data normalization is a conventional data processing method in the art, and this embodiment does not limit it;

[0151] S23, performing denoising processing on the second physiological parameter information to obtain pre-processed physiological parameter information.

[0152] Optionally, the performing denoising processing on the second physiological parameter information to obtain pre-processed physiological parameter information includes:

[0153] S231, processing the second physiological parameter information to obtain a real component and an imaginary component;

[0154] S232, performing wavelet transform on the real component to obtain first wavelet parameter information;

[0155] S233, performing wavelet transform on the imaginary component to obtain second wavelet parameter information;

[0156] S234, filtering the first wavelet parameter information to obtain third wavelet parameter information;

[0157] The filtering is low-pass filtering, and the parameters are set by experiments;

[0158] S235, filtering the second wavelet parameter information to obtain fourth wavelet parameter information;

[0159] The filtering is low-pass filtering, and the parameters are set by experiments;

[0160] S236, reconstructing the third wavelet parameter information and the fourth wavelet parameter information to obtain preprocessed physiological parameter information;

[0161] The reconstruction is wavelet reconstruction, which is the prior art in this field.

[0162] Optionally, the using the pre-processed physiological parameter information to train a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model includes:

[0163] S31, extracting features from the preprocessed physiological parameter information to obtain physiological parameter feature information; the physiological parameter feature information includes EEG feature parameter information, ECG feature parameter information, near infrared feature parameter information and blood feature parameter information;

[0164] S32, fusing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain fused characteristic parameter information;

[0165] S33, using the fused characteristic parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model.

[0166] Optionally, the extracting features of the preprocessed physiological parameter information to obtain physiological parameter feature information includes:

[0167] S311, extracting features from the preprocessed EEG information to obtain EEG feature parameter information;

[0168] S312, performing feature extraction on the preprocessed ECG information to obtain ECG feature parameter information;

[0169] S313, performing feature extraction on the pre-processed near infrared information to obtain near infrared feature parameter information;

[0170] S314, performing feature extraction on the pre-processed blood information to obtain blood feature parameter information.

[0171] Optionally, the step of extracting features from the preprocessed EEG information to obtain EEG feature parameter information includes:

[0172] S3111, extracting time-frequency features from the preprocessed EEG information to obtain time-frequency feature parameters;

[0173] The time-frequency characteristic parameter expression is:

[0174]

[0175] In the formula, S(m,n) is the time-frequency characteristic parameter, x(t) represents the preprocessed EEG information, g(i-nT) represents the window function, T represents the sliding window length, N represents the discrete Fourier transform length, m, n represent the time-frequency characteristic parameter frequency and time respectively, L represents the preprocessed EEG information length, and i represents the preprocessed EEG information sample point;

[0176] S3112, filtering the time-frequency characteristic parameters to obtain EEG characteristic parameter information, including:

[0177] Set the initial threshold δ th The target interval is [α, β], where 0≤α<β≤max{S(m,n)}, and S(m,n) is the time-frequency feature parameter;

[0178] Divide the target interval [α, β] into L equal parts;

[0179] The initial threshold δ th Set to α, according to The step size is incremented and iterated to obtain the threshold

[0180] Utilizing Thresholds Calculate the time-frequency characteristic parameters after filtering for each iteration

[0181]

[0182] Among them, S(m,n) represents the time-frequency characteristic parameters, represents the time-frequency characteristic parameters after filtering, a and b represent the time and frequency parameters in the threshold;

[0183] Calculate the difference between the time-frequency feature parameters obtained in each iteration and the previous iteration i1=1,2,…, i1 is the number of iterations;

[0184] The difference Perform curve fitting to obtain curve C. When curve C has an inflection point, the corresponding threshold is the optimal threshold.

[0185] The optimal threshold is used to perform optimal threshold filtering on the time-frequency feature parameter S(m,n) to obtain the filtered time-frequency feature parameter, which is the EEG feature parameter information.

[0186] Optionally, the performing feature extraction on the preprocessed ECG information to obtain ECG feature parameter information includes:

[0187] S3121, processing the pre-processed ECG information to obtain fuzzy parameter information; the fuzzy parameter information is expressed as:

[0188]

[0189] in, x(t) is the preprocessed ECG information, A x (θ,τ) is the fuzzy parameter information, t is the time variable, τ is the time shift variable, θ is the frequency variable corresponding to τ, and * indicates taking the conjugate;

[0190] S3122, performing Hadamard product calculation on the fuzzy parameter information and a preset kernel function to obtain ECG characteristic parameter information;

[0191] The Hadamard product calculation formula is:

[0192]

[0193] Among them, C x is the ECG characteristic parameter information, the integration range is -∞~∞, g(θ,τ) is the preset kernel function, a=0.001,β=50,u,t is the time variable, τ is the time shift variable, Ω is the frequency variable corresponding to u, and θ is the frequency variable corresponding to τ.

[0194] Optionally, the performing feature extraction on the pre-processed near infrared information to obtain near infrared feature parameter information includes:

[0195] S3131, using a near infrared information analysis model, performing time-frequency processing on the near infrared information to obtain first near infrared characteristic parameter information;

[0196] The near infrared information analysis model expression is:

[0197]

[0198] In the formula, t represents time, ω represents frequency, x() represents near-infrared information, τ represents delay, W(t,ω) is the first near-infrared characteristic parameter information, and * represents conjugation;

[0199] S3132, using a preset filtering model, filtering the first near infrared characteristic parameter information W(t,ω) to obtain near infrared characteristic parameter information

[0200] The preset filtering model expression is:

[0201]

[0202] Where W(τ,θ) represents the first near-infrared characteristic parameter information, W1(t-τ,ω-θ) is the two-dimensional template, and are two Gaussian functions, representing the spatial domain kernel and the pixel domain kernel respectively, τ is the time domain variable, and θ is the frequency domain variable;

[0203] and The specific expression is:

[0204]

[0205] In the formula, a and b represent the horizontal and vertical coordinates of the input pixel, i and j represent the coordinates of the center of the two-dimensional template, and σ s , σ r Represents the standard deviation of the Gaussian function.

[0206] Optionally, the extracting features of the pre-processed blood information to obtain blood feature parameter information includes:

[0207] S3141, performing matrix decomposition on the pre-processed blood information to obtain a left singular matrix, a singular value matrix, and a right singular matrix;

[0208] The matrix decomposition expression is:

[0209] Y=UAV

[0210] Among them, U is the left singular matrix, Y is the pre-processed blood information, A is the singular value matrix, V is the right singular matrix, U and V are both orthogonal matrices, and A is a diagonal matrix;

[0211] S3142, performing difference calculation on the left singular matrix and the right singular matrix to obtain a difference matrix;

[0212] S3143, performing cross-correlation calculation on the difference matrix and the singular value matrix to obtain blood characteristic parameter information.

[0213] Optionally, the fusing of the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information and the blood characteristic parameter information to obtain fused characteristic parameter information includes:

[0214] S3211, using a data centralization model, processing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain centralization parameter information;

[0215] The model expression in the data center is:

[0216]

[0217] The centralized parameter information is X=[w1-u, w2-u, w3-u, w4-u], w1 is EEG characteristic parameter information, w2 is ECG characteristic parameter information, w3 is near-infrared characteristic parameter information, w4 is blood characteristic parameter information, a1 is the weighting coefficient of w1, a2 is the weighting coefficient of w2, a3 is the weighting coefficient of w3, a4 is the weighting coefficient of w4, a1+a2+a3+a4=1;

[0218] Before centering, the four matrices w1, w2, w3 and w4 need to be processed by padding them with 0s to make the number of rows and columns of the four matrices equal for subsequent processing.

[0219] The calculation method of weight coefficients a1, a2, a3 and a4 is:

[0220] (1) The known indicator set w = {w1, w2, w3, w4};

[0221] (2) Calculate the correlation coefficient set R of the set w:

[0222]

[0223] where r ij ,i,j=1,2,…,4 is w i and w j The correlation coefficient of

[0224] (3) Solve the characteristic equation |λE-R|=0, where E is the unit matrix, and obtain the eigenvalues ​​λ1, λ2, λ3, and λ4;

[0225] (4) Calculate the cumulative contribution rate based on the eigenvalue:

[0226]

[0227] a i is the weight coefficient.

[0228] S3212, processing the centralization parameter information to obtain a covariance matrix;

[0229] The covariance matrix expression is:

[0230] ∑=XX T

[0231] Where ∑ is the covariance matrix and T is the transpose;

[0232] S3213, performing eigenvalue decomposition on the covariance matrix to obtain m eigenvalues;

[0233] S3214, sorting the m eigenvalues, and extracting eigenvectors corresponding to the first k eigenvalues;

[0234] S3215, using the feature vector reconstruction model, processing the feature vectors corresponding to the first k eigenvalues ​​to obtain fused feature parameter information;

[0235] The feature vector reconstruction model expression is:

[0236] Y=[u1X,u2X,…,u k X]

[0237] Among them, Y is the fusion feature parameter information, [u1,u2,…,u k ] is the eigenvector corresponding to the first k eigenvalues.

[0238] It can be seen that the present invention discloses a method for dynamic evaluation of physiological parameters, which includes: obtaining physiological parameter information of the user; mining and analyzing multiple physiological indicators to obtain dynamic evaluation results of physiological parameters. The method of the present invention only uses commonly used non-invasive parameters to dynamically evaluate the physiological parameters of the user, which reserves sufficient warning time for clinical intervention and can provide sufficient time for doctors to design treatment plans for patients.

[0239] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0240] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0241] Finally, it should be noted that the method for dynamic evaluation of physiological parameters disclosed in the embodiment of the present invention only discloses a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic evaluation of physiological parameters, characterized in that: The method comprises: S1, obtaining user physiological parameter information; the user physiological parameter information includes EEG information, ECG information, near infrared information and blood information; S2, preprocessing the user's physiological parameter information to obtain preprocessed physiological parameter information; the preprocessed physiological parameter information includes preprocessed EEG information, preprocessed ECG information, preprocessed near-infrared information and preprocessed blood information; S3, using the pre-processed physiological parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model; S4, using the optimized physiological parameter dynamic evaluation model, processing the user physiological parameter information to be processed to obtain a physiological parameter dynamic evaluation result.

2. The method for dynamic evaluation of physiological parameters according to claim 1, characterized in that: The preprocessing of the user's physiological parameter information to obtain preprocessed physiological parameter information includes: S21, performing abnormality check processing on the user's physiological parameter information to obtain first physiological parameter information; S22, normalizing the first physiological parameter information to obtain second physiological parameter information; S23, performing denoising processing on the second physiological parameter information to obtain pre-processed physiological parameter information.

3. The method for dynamic evaluation of physiological parameters according to claim 2, characterized in that: The performing denoising processing on the second physiological parameter information to obtain pre-processed physiological parameter information includes: S231, processing the second physiological parameter information to obtain a real component and an imaginary component; S232, performing wavelet transform on the real component to obtain first wavelet parameter information; S233, performing wavelet transform on the imaginary component to obtain second wavelet parameter information; S234, filtering the first wavelet parameter information to obtain third wavelet parameter information; S235, filtering the second wavelet parameter information to obtain fourth wavelet parameter information; S236, reconstructing the third wavelet parameter information and the fourth wavelet parameter information to obtain preprocessed physiological parameter information.

4. The method for dynamic evaluation of physiological parameters according to claim 1, characterized in that: The method of using the pre-processed physiological parameter information to train a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model includes: S31, extracting features from the preprocessed physiological parameter information to obtain physiological parameter feature information; the physiological parameter feature information includes EEG feature parameter information, ECG feature parameter information, near infrared feature parameter information and blood feature parameter information; S32, fusing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain fused characteristic parameter information; S33, using the fused characteristic parameter information, training a preset physiological parameter dynamic evaluation model to obtain an optimized physiological parameter dynamic evaluation model.

5. The method for dynamic evaluation of physiological parameters according to claim 4, characterized in that: The step of extracting features from the preprocessed physiological parameter information to obtain physiological parameter feature information includes: S311, extracting features from the preprocessed EEG information to obtain EEG feature parameter information; S312, performing feature extraction on the preprocessed ECG information to obtain ECG feature parameter information; S313, performing feature extraction on the pre-processed near infrared information to obtain near infrared feature parameter information; S314, performing feature extraction on the pre-processed blood information to obtain blood feature parameter information.

6. The method for dynamic evaluation of physiological parameters according to claim 5, characterized in that: The step of extracting features from the preprocessed EEG information to obtain EEG feature parameter information includes: S3111, extracting time-frequency features from the preprocessed EEG information to obtain time-frequency feature parameters; The time-frequency characteristic parameter expression is: In the formula, S(m, n) is the time-frequency characteristic parameter, x(t) represents the preprocessed EEG information, g(i-nT) represents the window function, T represents the sliding window length, N represents the discrete Fourier transform length, m and n represent the frequency and time of the time-frequency characteristic parameter respectively, L represents the length of the preprocessed EEG information, and i represents the preprocessed EEG information sample point; S3112, filtering the time-frequency characteristic parameters to obtain EEG characteristic parameter information, including: Set the initial threshold δ th The target interval is [α, β], where 0≤α<β≤max{S(m, n)}, and S(m, n) is the time-frequency feature parameter; Divide the target interval [α, β] into L equal parts; The initial threshold δ th Set to α, according to The step size is incremented and iterated to obtain the threshold Utilizing Thresholds Calculate the time-frequency characteristic parameters after filtering for each iteration Among them, S(m,n) represents the time-frequency characteristic parameters, represents the time-frequency characteristic parameters after filtering, a and b represent the time and frequency parameters in the threshold; Calculate the difference between the time-frequency feature parameters obtained in each iteration and the previous iteration i1=1, 2, …, i1 is the number of iterations; The difference Perform curve fitting to obtain curve c. When curve c has an inflection point, the corresponding threshold is the optimal threshold. The optimal threshold is used to perform optimal threshold filtering on the time-frequency feature parameter S(m, n) to obtain the filtered time-frequency feature parameter, which is the EEG feature parameter information.

7. The method for dynamic evaluation of physiological parameters according to claim 5, characterized in that: The step of extracting features from the preprocessed ECG information to obtain ECG feature parameter information includes: S3121, processing the pre-processed ECG information to obtain fuzzy parameter information; the fuzzy parameter information is expressed as: in, x(t) is the preprocessed ECG information, A x (θ, τ) is the fuzzy parameter information, t is the time variable, τ is the time shift variable, θ is the frequency variable corresponding to τ, and * indicates taking the conjugate; S3122, performing Hadamard product calculation on the fuzzy parameter information and a preset kernel function to obtain ECG characteristic parameter information; The Hadamard product calculation formula is: Among them, C x is the ECG characteristic parameter information, the integration range is -∞~∞, g(θ, τ) is the preset kernel function, a=0.001,β=50,u,t are time variables, τ is the time shift variable, Ω is the frequency variable corresponding to u, and θ is the frequency variable corresponding to τ.

8. The method for dynamic evaluation of physiological parameters according to claim 5, characterized in that: The step of extracting features from the pre-processed near infrared information to obtain near infrared feature parameter information includes: S3131, using a near infrared information analysis model, performing time-frequency processing on the near infrared information to obtain first near infrared characteristic parameter information; The near infrared information analysis model expression is: In the formula, t represents time, ω represents frequency, x() represents near-infrared information, τ represents delay, W(t, ω) is the first near-infrared characteristic parameter information, and * represents conjugation; S3132, using a preset filtering model, filtering the first near infrared characteristic parameter information W(t, ω) to obtain near infrared characteristic parameter information The preset filtering model expression is: Where W(τ, θ) represents the first near-infrared characteristic parameter information, W1(t-τ, ω-θ) is a two-dimensional template, and are two Gaussian functions, representing the spatial domain kernel and the pixel domain kernel respectively, τ is the time domain variable, and θ is the frequency domain variable; and The specific expression is: In the formula, a and b represent the horizontal and vertical coordinates of the input pixel, i and j represent the coordinates of the center of the two-dimensional template, and σ s , σ r Represents the standard deviation of the Gaussian function.

9. The method for dynamic evaluation of physiological parameters according to claim 5, characterized in that: The extracting characteristics of the pre-processed blood information to obtain blood characteristic parameter information includes: S3141, performing matrix decomposition on the pre-processed blood information to obtain a left singular matrix, a singular value matrix, and a right singular matrix; The matrix decomposition expression is: Y=UAV Among them, U is the left singular matrix, Y is the pre-processed blood information, A is the singular value matrix, V is the right singular matrix, U and V are both orthogonal matrices, and A is a diagonal matrix; S3142, performing difference calculation on the left singular matrix and the right singular matrix to obtain a difference matrix; S3143, performing cross-correlation calculation on the difference matrix and the singular value matrix to obtain blood characteristic parameter information.

10. The method for dynamic evaluation of physiological parameters according to claim 4, characterized in that: The fusing of the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information and the blood characteristic parameter information to obtain fused characteristic parameter information includes: S3211, using a data centralization model, processing the EEG characteristic parameter information, the ECG characteristic parameter information, the near-infrared characteristic parameter information, and the blood characteristic parameter information to obtain centralization parameter information; The model expression in the data center is: The centralized parameter information is X=[w1-u, w2-u, w3-u, w4-u], w1 is EEG characteristic parameter information, w2 is ECG characteristic parameter information, w3 is near-infrared characteristic parameter information, w4 is blood characteristic parameter information, a1 is the weighting coefficient of w1, a2 is the weighting coefficient of w2, a3 is the weighting coefficient of w3, a4 is the weighting coefficient of w4, a1+a2+a3+a4=1; S3212, processing the centralization parameter information to obtain a covariance matrix; The covariance matrix expression is: ∑=XX T Where ∑ is the covariance matrix and T is the transpose; S3213, performing eigenvalue decomposition on the covariance matrix to obtain m eigenvalues; S3214, sorting the m eigenvalues, and extracting eigenvectors corresponding to the first k eigenvalues; S3215, using the feature vector reconstruction model, processing the feature vectors corresponding to the first k eigenvalues ​​to obtain fused feature parameter information; The feature vector reconstruction model expression is: Y=[u1X,u2X,…,u k X] Among them, Y is the fusion feature parameter information, [u1, u2, ..., u k ] is the eigenvector corresponding to the first k eigenvalues.

Images