A method and system for predicting cerebral hemorrhage based on PCA-BP neural network

Through the PCA-BP neural network combined with BP algorithm, a cerebral hemorrhage and arm amplitude detection model was established, which solved the problem of low accuracy in the prediction of cerebral hemorrhage in the existing technology, and achieved accurate judgment and timely alarms for cerebral hemorrhage, ensuring the best Good rescue time arrest.

CN114795145BActive Publication Date: 2024-11-26SICHUAN UNIVERSITY OF SCIENCE AND ENGINEERING

Patent Information

Application Number
CN202210359458.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-06
Publication Date
2024-11-26
Estimated Expiration
2042-04-06

AI Technical Summary

Technical Problem

The prior art has low accuracy when predicting cerebral hemorrhage, and cannot effectively judge the symptoms of hemiplegia, resulting in the possibility of missing the best rescue time.

Method used

Using the cerebral hemorrhage prediction method based on PCA-BP neural network, data is collected through physiological information collection module and arm amplitude detection module, a cerebral hemorrhage probability prediction model and arm amplitude detection model are established, and dimensionality reduction and prediction are combined with the BP neural network algorithm to achieve dimensionality reduction and prediction, so as to achieve the Accurate judgment of cerebral hemorrhage.

Benefits of technology

It significantly improves the accuracy of prediction of cerebral hemorrhage, can promptly call the alarm and provide help information, especially when the arm amplitude is abnormal, it can delay the alarm to ensure timely rescue.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114795145B_ABST
    Figure CN114795145B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for predicting cerebral hemorrhage based on a PCA-BP neural network, the method comprising the following steps: respectively collecting human physiological data and arm amplitude data through a physiological information acquisition module and an arm amplitude detection module as inputs to a cerebral hemorrhage probability prediction model and an arm amplitude detection model; establishing a cerebral hemorrhage probability prediction model and an arm amplitude detection model through a PCA-BP neural network algorithm; judging whether a cerebral hemorrhage occurs in a detected human body through the cerebral hemorrhage probability prediction model, and judging whether a detected human body is paralyzed through the arm amplitude detection model. The present invention establishes a cerebral hemorrhage probability prediction model and an arm amplitude detection model through a PCA-BP neural network algorithm, and predicts cerebral hemorrhage in combination with the cerebral hemorrhage probability prediction model and the arm amplitude detection model, which greatly improves the accurate judgment of cerebral hemorrhage in a human body.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method and system for predicting cerebral hemorrhage based on a PCA-BP neural network. Background Art

[0002] Cerebral hemorrhage refers to bleeding caused by non-traumatic rupture of blood vessels in the brain parenchyma, accounting for 20% to 30% of all strokes, with an acute mortality rate of 30% to 40%. The cause of occurrence is mainly related to cerebrovascular lesions, that is, closely related to hyperlipidemia, diabetes, hypertension, vascular aging, smoking, etc. Patients with cerebral hemorrhage often develop the disease suddenly due to emotional excitement or strenuous effort, and the early mortality rate is very high. Most of the survivors have sequelae such as movement disorders, cognitive disorders, speech and swallowing disorders to varying degrees.

[0003] At present, the judgment mechanism for predicting cerebral hemorrhage in the market is to judge that the wearer has cerebral hemorrhage when one or more physiological indicators (such as pulse, blood pressure, etc.) exceed the set threshold, or when the wearer faints. When a sudden cerebral hemorrhage occurs, most people will have symptoms of hemiplegia (such as arm paralysis, etc.), but the market does not have a model to judge the occurrence of hemiplegia, and it cannot be combined with the threshold judgment model. Therefore, when a simple threshold judgment is made through abnormal physiological indicators, the accuracy of predicting cerebral hemorrhage is not high, which may cause the wearer to miss the "golden time" for the best rescue. Summary of the invention

[0004] One of the purposes of the present invention is to provide a method for predicting cerebral hemorrhage based on a PCA-BP neural network to solve the technical problem of how to improve the accuracy of cerebral hemorrhage prediction.

[0005] The first object of the present invention is achieved by adopting the following technical scheme: a method for predicting cerebral hemorrhage based on PCA-BP neural network, comprising the following steps:

[0006] The physiological information acquisition module collects human physiological data, and stores the collected physiological data in the database of the main control module as input to the cerebral hemorrhage probability prediction model;

[0007] The arm amplitude detection module collects the human body's arm amplitude data, and stores the collected arm amplitude data in the database of the main control module as the input of the arm amplitude detection model;

[0008] The PCA-BP neural network algorithm was used to establish a cerebral hemorrhage probability prediction model and an arm amplitude detection model;

[0009] The cerebral hemorrhage probability prediction model predicts the probability of cerebral hemorrhage in the human body based on physiological data. When the probability of cerebral hemorrhage in the human body exceeds the preset threshold, an early warning will be issued; at the same time, the arm amplitude detection model detects the human arm amplitude data. When the arm amplitude is detected to be abnormal, an alarm will be issued immediately for help; if the arm amplitude is detected to be normal but cerebral hemorrhage is about to occur, a delayed alarm will be issued for help.

[0010] Furthermore, the physiological information acquisition module includes a blood pressure and pulse detection module and a body temperature detection module, and the blood pressure and pulse detection module and the body temperature detection module are both connected to the main control module to collect human blood pressure, pulse and body temperature data, and store them in the database of the main control module as input to the cerebral hemorrhage prediction model.

[0011] Furthermore, the arm amplitude detection module includes a three-axis gravity acceleration sensor and an electronic gyroscope, which are used to collect the shaking speed, angle and angular acceleration of the human arm, and store them in the database of the main control module as input of the arm amplitude detection model.

[0012] Furthermore, the PCA-BP neural network algorithm is used to establish a cerebral hemorrhage probability prediction model, which includes the following steps:

[0013] The principal component analysis (PCA) is used to reduce the 56-dimensional input variables to 12 dimensions, and then the probability prediction is performed through the BP neural network, which includes three layers of neurons: input layer, hidden layer and output layer.

[0014] The input signal enters the network from the input layer, passes through the intermediate hidden layer and moves forward to the output layer, where the actual output result corresponding to the input is obtained;

[0015] If the actual output of the network does not match the expected output, the error back propagation phase begins;

[0016] The above process is repeated so that the connection weights between nodes in each layer of the network are continuously adjusted to complete the learning process of the BP neural network.

[0017] Complete the establishment of the probability prediction model for cerebral hemorrhage.

[0018] Furthermore, the method steps for establishing the arm amplitude detection model through the PCA-BP neural network algorithm are consistent with the method steps for establishing the cerebral hemorrhage probability prediction model through the PCA-BP neural network algorithm.

[0019] Furthermore, the method for predicting the probability of cerebral hemorrhage in a human body based on physiological data by using the cerebral hemorrhage probability prediction model includes the following steps:

[0020] The physiological data is used as the input of PCA, and the input after dimension reduction is obtained through PCA, and then used as the input of BP neural network, and enters the hidden layer through the input layer of BP neural network, and then the output of the hidden layer is sent to the output layer;

[0021] After the operation of the input layer, hidden layer and output layer of the BP neural network, the predicted value of the probability of cerebral hemorrhage is obtained;

[0022] Finally, the output layer outputs the standardized cerebral hemorrhage probability prediction value, which is compared with the cerebral hemorrhage preset threshold. If it is higher than or equal to the cerebral hemorrhage preset threshold, it immediately enters the next alarm link; if it is lower than the cerebral hemorrhage preset threshold, it returns to daily detection;

[0023] Among them, the mathematical expressions of the hidden layer and the output layer are as follows:

[0024] Hidden layer 1

[0025] Hidden layer 2

[0026] Hidden layer third layer

[0027] Output Layer

[0028] In the formula, h j 、h k 、h l is the output of the hidden layer neurons; y is the output of the output layer; w i1 ,w j2 ,w k3 ,w lo are the input weights of the hidden layer and the output layer, respectively, satisfying b is the bias of the neuron; X = x1, x2, …, x n ] T It is the input of BP neural network and the input of the first layer of neurons in the hidden layer, that is, the 12-dimensional physiological index data after PCA dimensionality reduction processing; f1, f2, f3, f o They are the activation functions of the neurons in the three hidden layers and the output layer. The activation function of the neurons in the first layer of the hidden layer is tan sig, which can be mapped between -1 and 1. Its expression is: f1x=tansig(x)=2 / (1+exp(-2*x))-1; the activation function of the neurons in the second layer of the hidden layer is softmax, which maps the output values ​​of multiple neurons to (0,1), and the sum is 1. When outputting, the one with the largest value is selected as the predicted value. Its expression is: The activation functions of the third layer of the hidden layer and the output layer neurons are both log sig, and the output value is between 0 and 1, so that the network outputs a specific probability prediction value, and its expression is: f3x = f o (x)=logsig(x)=1 / (1+exp(-x)).

[0029] The second purpose of the present invention is to provide a cerebral hemorrhage prediction system based on PCA-BP neural network to solve the technical problem of how to improve the accuracy of cerebral hemorrhage prediction.

[0030] The second objective of the present invention is achieved through the following technical solutions: a cerebral hemorrhage prediction system based on a PCA-BP neural network, comprising a physiological information acquisition module, an arm amplitude detection module and a main control module, wherein the physiological information acquisition module is used to collect human physiological data, and the collected physiological data are stored in a database of the main control module, the arm amplitude detection module is used to collect human arm amplitude data, and the collected arm amplitude data are stored in a database of the main control module, the cerebral hemorrhage probability prediction model and the arm amplitude detection model in the main control module are used to analyze the physiological data and the arm amplitude data respectively to determine whether cerebral hemorrhage occurs in the human body, and an alarm message and a distress message are sent through an alarm module and a distress module connected to the main control module; the system also comprises a communication positioning module and a display module, and the communication positioning module and the display module are both connected to the main control module.

[0031] The beneficial effect of the present invention is that a cerebral hemorrhage probability prediction model and an arm amplitude detection model are established through the PCA-BP neural network algorithm, and cerebral hemorrhage is predicted by combining the cerebral hemorrhage probability prediction model and the arm amplitude detection model, which greatly improves the accurate judgment of human cerebral hemorrhage. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 the structures shown in these drawings without paying creative work.

[0033] Figure 1 It is a flow chart of the present invention;

[0034] Figure 2 It is a schematic diagram of the working principle of the present invention;

[0035] Figure 3 is a system block diagram of the present invention;

[0036] Figure 4 It is a prediction model based on PCA-BP neural network algorithm;

[0037] Figure 5 This is a schematic diagram of the structure of a smart bracelet;

[0038] In the figure, 100 is the bracelet body, 101 is the first button, 102 is the inside of the bracelet, 103 is the second button, 104 is the blood pressure and pulse detection module, 105 is the infrared temperature measurement module, and 106 is the bracelet strap. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions 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 part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0040] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0041] Embodiment 1:

[0042] See also Figure 1 , a method for predicting cerebral hemorrhage based on PCA-BP neural network, comprising the following steps:

[0043] The physiological information acquisition module collects human physiological data, and stores the collected physiological data in the database of the main control module as input to the cerebral hemorrhage probability prediction model;

[0044] The arm amplitude detection module collects the human body's arm amplitude data, and stores the collected arm amplitude data in the database of the main control module as the input of the arm amplitude detection model;

[0045] The PCA-BP neural network algorithm was used to establish a cerebral hemorrhage probability prediction model and an arm amplitude detection model;

[0046] The cerebral hemorrhage probability prediction model predicts the probability of cerebral hemorrhage in the human body based on physiological data. When the probability of cerebral hemorrhage in the human body exceeds the preset threshold, an early warning will be issued; at the same time, the arm amplitude detection model detects the human arm amplitude data. When the arm amplitude is detected to be abnormal, an alarm will be issued immediately for help; if the arm amplitude is detected to be normal but cerebral hemorrhage is about to occur, a delayed alarm will be issued for help.

[0047] In this embodiment, the physiological information acquisition module includes a blood pressure and pulse detection module and a body temperature detection module. Both the blood pressure and pulse detection module and the body temperature detection module are connected to the main control module to collect human blood pressure, pulse and body temperature data, and store them in the database of the main control module as input to the cerebral hemorrhage prediction model.

[0048] For further information, see Figure 2 The main control module receives and analyzes the arm vibration amplitude, body temperature, blood pressure, and pulse data from the arm amplitude detection module, body temperature detection module, and blood pressure and pulse detection module, and analyzes the above data through the cerebral hemorrhage probability prediction model and the arm amplitude detection model in the main control module. When abnormal physiological indicators such as blood pressure, pulse, and body temperature are detected, the detected person is warned. The arm amplitude detection model determines whether the arm amplitude of the detected person is abnormal. If abnormal, the alarm module immediately sends out positioning information and a distress signal. If the detected person finds that his or her physiological indicators are abnormal and has symptoms of cerebral hemorrhage, the distress module can be activated to send a distress signal, so that the alarm module sends out positioning information and a distress signal.

[0049] In this embodiment, the arm amplitude detection module includes a three-axis gravity acceleration sensor and an electronic gyroscope, which are used to collect the shaking speed, angle and angular acceleration of the human arm, and store them in the database of the main control module as input of the arm amplitude detection model.

[0050] In this embodiment, the method for establishing a cerebral hemorrhage probability prediction model by using a PCA-BP neural network algorithm includes the following steps:

[0051] The principal component analysis (PCA) is used to reduce the 56-dimensional input variables to 12 dimensions, and then the probability prediction is performed through the BP neural network, which includes three layers of neurons: input layer, hidden layer and output layer.

[0052] The input signal enters the network from the input layer, passes through the intermediate hidden layer and moves forward to the output layer, where the actual output result corresponding to the input is obtained;

[0053] If the actual output of the network does not match the expected output, the error back propagation phase begins; back propagation is the process of adjusting the connection weights between nodes in each layer of the network.

[0054] The above process is repeated so that the connection weights between nodes in each layer of the network are continuously adjusted to complete the learning process of the BP neural network.

[0055] Complete the establishment of the probability prediction model for cerebral hemorrhage.

[0056] In this embodiment, the method steps of establishing the arm amplitude detection model through the PCA-BP neural network algorithm are consistent with the method steps of establishing the cerebral hemorrhage probability prediction model through the PCA-BP neural network algorithm.

[0057] In this embodiment, the method for predicting the probability of cerebral hemorrhage in the human body based on physiological data by using the cerebral hemorrhage probability prediction model is specifically described in Figure 4 , the method comprises the following steps:

[0058] The physiological data is used as the input of PCA, and the input after dimension reduction is obtained through PCA, and then used as the input of BP neural network, and enters the hidden layer through the input layer of BP neural network, and then the output of the hidden layer is sent to the output layer;

[0059] After the operation of the input layer, hidden layer and output layer of the BP neural network, the predicted value of the probability of cerebral hemorrhage is obtained;

[0060] Finally, the output layer outputs the standardized cerebral hemorrhage probability prediction value, which is compared with the cerebral hemorrhage preset threshold (preferably, after debugging, the cerebral hemorrhage preset threshold is set at 0.82). If it is higher than or equal to the cerebral hemorrhage preset threshold, it immediately enters the next alarm link; if it is lower than the cerebral hemorrhage preset threshold, it returns to daily detection;

[0061] Among them, the mathematical expressions of the hidden layer and the output layer are as follows:

[0062] Hidden layer 1

[0063] Hidden layer 2

[0064] Hidden layer third layer

[0065] Output Layer

[0066] In the formula, h j 、h k 、h l is the output of the hidden layer neurons; y is the output of the output layer; w i1 ,w j2 ,w k3 ,w lo are the input weights of the hidden layer and the output layer, respectively, satisfying b is the bias of the neuron; X = x1, x2, …, x n ] T It is the input of BP neural network and the input of the first layer of neurons in the hidden layer, that is, the 12-dimensional physiological index data after PCA dimensionality reduction processing; f1, f2, f3, f oThey are the activation functions of the neurons in the three hidden layers and the output layer. The activation function of the neurons in the first layer of the hidden layer is tan sig, which can be mapped between -1 and 1. Its expression is: f1x=tansig(x)=2 / (1+exp(-2*x))-1; the activation function of the neurons in the second layer of the hidden layer is softmax, which maps the output values ​​of multiple neurons to (0,1), and the sum is 1. When outputting, the one with the largest value is selected as the predicted value. Its expression is: The activation functions of the third layer of the hidden layer and the output layer neurons are both log sig, and the output value is between 0 and 1, so that the network outputs a specific probability prediction value, and its expression is: f3(x) = f o (x)=logsig(x)=1 / (1+exp(-x)).

[0067] After the main control module obtains different physiological data from multiple modules (including the body temperature detection module and the blood pressure and pulse detection module), the cerebral hemorrhage probability prediction model learns and trains the data. When the physiological data of the human body, namely body temperature, blood pressure, and pulse, are different from the state in daily life, the cerebral hemorrhage probability prediction model predicts the probability of cerebral hemorrhage in the wearer. When the wearer's cerebral hemorrhage probability exceeds the preset probability threshold (0.82), an early warning will be issued through the alarm module, and the arm amplitude detection model will detect the amplitude of the human arm. If the human arm is paralyzed, the emergency module will immediately alarm for help; if the human arm amplitude is judged to be normal at this time but the human body is about to or is experiencing cerebral hemorrhage, a delayed alarm will be issued for help. The arm amplitude detection model detects the arm amplitude by detecting the shaking angular velocity, angle and angular acceleration of the human arm. When a sudden cerebral hemorrhage occurs, some people will experience hemiplegia due to blood compressing the motor area of ​​the brain. At this time, the human body is no longer able to manually alarm, and the main control module will automatically detect the arm's movement angle, angular velocity and angular acceleration in three dimensions (i.e., x, y, and z axes). When the main control module detects and determines that the characteristics of the human body's angle, angular velocity and angular acceleration meet the characteristics of arm hemiplegia, the main control module will combine other data (such as the probability value of cerebral hemorrhage) to determine whether the human body has a sudden cerebral hemorrhage. Therefore, it is necessary to detect the angle, angular velocity and angular velocity of the wrist in three dimensions at different times. At the same time, because the hardware equipment samples the arm amplitude information ten times per second, considering the urgent rescue time, 10 seconds is selected as the sampling duration to obtain 900 data.

[0068] The arm amplitude detection model detects the human arm amplitude data: the arm amplitude detection model is similar to the cerebral hemorrhage probability prediction model, and still uses the PCA-BP neural network algorithm. The input is the data of the angle, angular velocity and angular acceleration of the arm in the three-dimensional direction sampled for 10 seconds. The final output is the arm amplitude detection value (the output range is between 0 and 1). The arm amplitude detection value is compared with the arm amplitude preset threshold (preferably, after debugging, the arm amplitude preset threshold is set to 0.85). If it is higher than or equal to the threshold, it will immediately enter the emergency rescue link (send a distress alarm message through the distress module); if it is lower than the threshold, a delayed alarm will be issued. The combination of the cerebral hemorrhage probability prediction model and the arm amplitude detection model solves the problem of inaccurate current cerebral hemorrhage detection bracelets.

[0069] In addition, it should be noted that the parameters used in the prediction model for the probability of cerebral hemorrhage in this embodiment include pulse, blood pressure, body temperature and medical history related to the human body. However, those who are familiar with this field should understand that other parameters can also be obtained by using sensors to combine and train the BP neural network to obtain a better prediction model.

[0070] For further information, see Figure 3 A cerebral hemorrhage prediction system based on PCA-BP neural network includes a physiological information acquisition module, an arm amplitude detection module and a main control module. The physiological information acquisition module is used to collect human physiological data, and the collected physiological data are stored in the database of the main control module. The arm amplitude detection module is used to collect human arm amplitude data, and the collected arm amplitude data are stored in the database of the main control module. The physiological data and the arm amplitude data are analyzed respectively by the cerebral hemorrhage probability prediction model and the arm amplitude detection model in the main control module to determine whether cerebral hemorrhage occurs in the human body, and an alarm message and a distress message are sent through the alarm module and the distress module connected to the main control module; it also includes a communication positioning module and a display module, and the communication positioning module and the display module are both connected to the main control module.

[0071] In this embodiment, the physiological information acquisition module includes a blood pressure and pulse detection module and a body temperature detection module, which respectively detect the blood pressure, pulse and body temperature of the human body. The arm amplitude detection module includes a three-axis gravity acceleration sensor and an electronic gyroscope, which are used to detect the degree of arm amplitude; the communication positioning module locates the position of the detected human body and sends a distress message when a cerebral hemorrhage occurs in the human body; the alarm module allows the detected person to ask for help when he finds that his body is abnormal; the distress module is used to send a distress signal; the display module is used to display time and related physiological information.

[0072] For further information, see Figure 5The present invention provides a smart bracelet, which uses the above-mentioned PCA-BP neural network-based cerebral hemorrhage prediction method to perform real-time detection on a wearer. The smart bracelet includes a bracelet body 100, and the bracelet body 100 is provided with an arm amplitude detection module, a body temperature detection module, a blood pressure pulse detection module, a communication positioning module, a display module, an alarm module and a help module. The arm amplitude detection module, the body temperature detection module, the blood pressure pulse detection module, the communication positioning module, the display module, the alarm module and the help module are all connected to a main control module through a signal line. The main control module receives signals from the arm amplitude detection module, the body temperature detection module, the blood pressure pulse detection module and the communication positioning module, and judges and alarms whether the wearer has cerebral hemorrhage through a cerebral hemorrhage probability prediction model and an arm amplitude detection model provided in the main control module. It can be understood that the wearer's blood pressure, pulse and body temperature are detected respectively by the blood pressure and pulse detection module and the body temperature detection module. The arm amplitude detection module includes a three-axis gravity acceleration sensor and an electronic gyroscope to detect the degree of arm amplitude. The communication positioning module locates the position of the detected human body and sends out a distress message when a cerebral hemorrhage occurs in the human body; the alarm module allows the detected person to ask for help when he finds that his body is abnormal; the distress module is used to send a distress signal; the display module is used to display time and related physiological information.

[0073] In this embodiment, the smart bracelet is used to be worn on the wearer's hand. The bracelet strap 106 can be adjusted according to the size of the wrist to facilitate wearing by different groups of people. The first button 101 and the second button 103 are arranged on the side of the dial, connected to the alarm module and the rescue module inside the bracelet 102, and provide input signals. The front of the bracelet body 100 is provided with an ISP screen, which is connected to the main control module to display the current bracelet status information. The alarm module and the communication positioning module are integrated inside the bracelet 102 and controlled by the main control module. The alarm module is connected to a speaker and an LED light, which flashes and emits an alarm sound when receiving an alarm signal. The bracelet is also provided with a motion sensor inside 102, including a three-axis acceleration and an electronic gyroscope module, which transmits data to the main control module. The main control module runs an arm amplitude detection model to determine whether the wearer is paralyzed. The physiological data acquisition module consists of a blood pressure pulse module detection 104 and an infrared temperature measurement module 105. The module detects the wearer's blood pressure, pulse and body temperature in a non-invasive manner, and then sends them to the main control module through the UART1 serial port of the stm32f103c8. The main control module stores it in the database in the bracelet. When the main control module runs the cerebral hemorrhage probability prediction model, it reads the historical data on the bracelet as input to determine whether the probability of cerebral hemorrhage in the wearer exceeds the threshold, and then issues an early warning.

[0074] Furthermore, the alarm module includes two warning modes: the first one is set on the bracelet body 100, and in an emergency, a call can be made to a preset number in the smart bracelet, and the wearer's current location and distress information can be sent to the emergency center; the second one is that the smart bracelet will sound an alarm and the LED light on the bracelet will flash. It can be understood that the main control module is the control center of the smart bracelet, and other modules are connected to the main control module to provide data to the main control module. The data that the main control module can obtain include: blood pressure, pulse, body temperature, and arm amplitude related data (three-axis acceleration, angular velocity and angular acceleration). According to these data inputs, the main control module uses the cerebral hemorrhage probability prediction model and arm amplitude detection model established by the PCA-BP neural network algorithm to judge cerebral hemorrhage, and then control the output. The main control module has two mechanisms for judgment: first, it uses the cerebral hemorrhage probability prediction model obtained through historical physiological data to predict the probability of cerebral hemorrhage. When the algorithm determines that the wearer's cerebral hemorrhage probability exceeds the preset probability threshold, the main control module controls the alarm module to perform related alarm operations; second, the arm amplitude detection module obtains the three-axis acceleration, angular velocity and angular acceleration, runs the arm amplitude detection model, and determines whether the wearer is currently paralyzed. If the wearer is detected to be paralyzed, the alarm module is controlled to sound the alarm and flash the LED. If it is a false alarm, the wearer can actively press the button on the bracelet to turn it off. If the alarm is not turned off within the preset time of the main control module, it is determined that the wearer has cerebral hemorrhage, and the preset number is notified, the emergency number is dialed, and the positioning information is sent. The main control module is built into the bracelet body 100, and the arm amplitude detection module, physiological information collection module, alarm module and rescue module are respectively connected to the main control module and are also placed inside the bracelet shell. The physiological information collection module is placed on the side that fits the skin. The body temperature detection module and blood pressure pulse detection module in the physiological information collection module use the principles of infrared temperature measurement and PPG photoelectric capacitance method to perform non-invasive detection of the wearer's body temperature, pulse and blood pressure data.

[0075] The present invention establishes a cerebral hemorrhage probability prediction model and an arm amplitude detection model through the PCA-BP neural network algorithm, and predicts cerebral hemorrhage in combination with the cerebral hemorrhage probability prediction model and the arm amplitude detection model, which greatly improves the accurate judgment of human cerebral hemorrhage.

[0076] It should be noted that, for the above embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification belong to preferred embodiments, and the actions involved are not necessarily required by the present application.

[0077] In addition, the terms "connected" and "set" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "connected" or "set" may explicitly or implicitly include one or more of the features. Moreover, the terms "connected", "set", etc. are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate so that the embodiments of the invention described herein can be implemented in an order other than those illustrated or described herein.

[0078] The above embodiments describe the basic principles and main features of the present invention and the advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the changes and modifications made by those skilled in the art shall be within the scope of protection of the appended claims of the present invention without departing from the spirit and scope of the present invention.

Claims

1. A method for predicting cerebral hemorrhage based on PCA-BP neural network, characterized in that: The steps include: The physiological information acquisition module collects human physiological data, and stores the collected physiological data in the database of the main control module as input to the cerebral hemorrhage probability prediction model; The arm amplitude detection module collects the human body's arm amplitude data, and stores the collected arm amplitude data in the database of the main control module as the input of the arm amplitude detection model; The PCA-BP neural network algorithm was used to establish a cerebral hemorrhage probability prediction model and an arm amplitude detection model; The cerebral hemorrhage probability prediction model predicts the probability of cerebral hemorrhage in the human body based on physiological data. When the probability of cerebral hemorrhage in the human body exceeds the preset threshold, an early warning will be issued; at the same time, the arm amplitude detection model detects the human arm amplitude data. When the arm amplitude is detected to be abnormal, an alarm will be immediately issued for help; if the arm amplitude is detected to be normal but cerebral hemorrhage is about to occur, a delayed alarm will be issued for help; Establishing a cerebral hemorrhage probability prediction model using the PCA-BP neural network algorithm includes the following steps: The principal component analysis (PCA) is used to reduce the 56-dimensional input variables to 12 dimensions, and then the probability prediction is performed through the BP neural network, which includes three layers of neurons: input layer, hidden layer and output layer. The input signal enters the network from the input layer, passes through the intermediate hidden layer and moves forward to the output layer, where the actual output result corresponding to the input is obtained; If the actual output of the network does not match the expected output, the error back propagation phase begins; The above process is repeated so that the connection weights between nodes in each layer of the network are continuously adjusted to complete the learning process of the BP neural network. Complete the establishment of a prediction model for the probability of cerebral hemorrhage; The method steps for establishing the arm amplitude detection model through the PCA-BP neural network algorithm are consistent with the method steps for establishing the cerebral hemorrhage probability prediction model through the PCA-BP neural network algorithm.

2. A method for predicting cerebral hemorrhage based on PCA-BP neural network as claimed in claim 1, characterized in that: The physiological information acquisition module includes a blood pressure and pulse detection module and a body temperature detection module. Both the blood pressure and pulse detection module and the body temperature detection module are connected to the main control module to collect human blood pressure, pulse and body temperature data, and store them in the database of the main control module as input to the cerebral hemorrhage prediction model.

3. The method for predicting cerebral hemorrhage based on PCA-BP neural network as claimed in claim 1, characterized in that: The arm amplitude detection module includes a three-axis gravity acceleration sensor and an electronic gyroscope, which are used to collect the shaking speed, angle and angular acceleration of the human arm, and store them in the database of the main control module as input of the arm amplitude detection model.

4. The method for predicting cerebral hemorrhage based on PCA-BP neural network as claimed in claim 1, characterized in that: The method for predicting the probability of cerebral hemorrhage in a human body based on physiological data by using a cerebral hemorrhage probability prediction model includes the following steps: The physiological data is used as the input of PCA, and the input after dimension reduction is obtained through PCA, and then used as the input of BP neural network, and enters the hidden layer through the input layer of BP neural network, and then the output of the hidden layer is sent to the output layer; After the operation of the input layer, hidden layer and output layer of the BP neural network, the predicted value of the probability of cerebral hemorrhage is obtained; Finally, the output layer outputs the standardized cerebral hemorrhage probability prediction value, which is compared with the cerebral hemorrhage preset threshold. If it is higher than or equal to the cerebral hemorrhage preset threshold, it immediately enters the next alarm link; if it is lower than the cerebral hemorrhage preset threshold, it returns to daily detection; Among them, the mathematical expressions of the hidden layer and the output layer are as follows: Hidden layer 1 Hidden layer 2 Hidden layer third layer Output Layer In the formula, h j 、h k 、h l is the output of the hidden layer neurons; y is the output of the output layer; w i1 ,w j2 ,w k3 ,w lo are the input weights of the hidden layer and the output layer, respectively, satisfying b is the bias of the neuron; X = [x1, x2, ..., x n ] T It is the input of BP neural network and the input of the first layer of neurons in the hidden layer, that is, the 12-dimensional physiological index data after PCA dimensionality reduction processing; f1, f2, f3, f o They are the activation functions of the neurons in the three hidden layers and the output layer. The activation function of the neurons in the first layer of the hidden layer is tansig, which can be mapped between -1 and 1. Its expression is: f1(x)=tansig(x)=2 / (1+exp(-2*x))-1; the activation function of the neurons in the second layer of the hidden layer is softmax, which maps the output values ​​of multiple neurons to (0,1), and the sum is 1. When outputting, the largest value is selected as the predicted value. Its expression is: The activation functions of the third layer of the hidden layer and the output layer neurons are both logsig, and the output value is between 0 and 1, so that the network outputs a specific probability prediction value, and its expression is: f3(x) = f o (x)=logsig(x)=1 / (1+exp(-x)).

5. A system for predicting cerebral hemorrhage based on PCA-BP neural network, used to implement a method for predicting cerebral hemorrhage based on PCA-BP neural network as claimed in any one of claims 1 to 4, characterized in that: It includes a physiological information acquisition module, an arm amplitude detection module and a main control module. The physiological information acquisition module is used to collect human physiological data, and the collected physiological data are stored in the database of the main control module. The arm amplitude detection module is used to collect human arm amplitude data, and the collected arm amplitude data are stored in the database of the main control module. The cerebral hemorrhage probability prediction model and the arm amplitude detection model in the main control module are used to analyze the physiological data and the arm amplitude data respectively to determine whether cerebral hemorrhage occurs in the human body, and an alarm message and a distress message are sent out through the alarm module and the distress module connected to the main control module; it also includes a communication positioning module and a display module, and the communication positioning module and the display module are both connected to the main control module.

Citation Information

Patent Citations

  • Health detector

    CN105167763A

  • Monitoring system for stroke people

    CN109741834A

  • Hematoma enlargement risk prediction system and method for cerebral hemorrhage patient

    CN112991320A

  • Portable cerebral hemorrhage prediction bracelet

    CN217186102U

  • Systems and methods for diagnosing a stroke condition

    US20220044821A1

Cited By

  • Cerebral hemorrhage risk prediction method and system based on multi-modal learning

    CN121662369A