CN120203550A - An arrhythmia diagnosis system based on multi-sensor fusion - Google Patents
An arrhythmia diagnosis system based on multi-sensor fusion Download PDFInfo
- Publication number
- CN120203550A CN120203550A CN202510421947.8A CN202510421947A CN120203550A CN 120203550 A CN120203550 A CN 120203550A CN 202510421947 A CN202510421947 A CN 202510421947A CN 120203550 A CN120203550 A CN 120203550A
- Authority
- CN
- China
- Prior art keywords
- arrhythmia
- signal
- module
- algorithm
- input signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
- A61B5/024—Measuring pulse rate or heart rate
- A61B5/02405—Determining heart rate variability
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
- A61B5/024—Measuring pulse rate or heart rate
- A61B5/02444—Details of sensor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
- G06F18/256—Fusion techniques of classification results, e.g. of results related to same input data of results relating to different input data, e.g. multimodal recognition
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Theoretical Computer Science (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Cardiology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Physiology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Heart & Thoracic Surgery (AREA)
- General Physics & Mathematics (AREA)
- Veterinary Medicine (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Fuzzy Systems (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
本发明涉及智能医疗技术领域，尤其是一种基于多传感器融合的心律失常诊断系统，所述系统包括：心电信号模块，包括心电信号传感器，用于采集心电信号；脉搏波信号模块，包括脉搏波传感器，用于采集脉搏波信号；三轴加速度模块，包括加速度传感器，用于采集加速度信号；主控芯片模块，作为数据处理模块，用于数据接收和数据分析，得到心律失常诊断结果；蓝牙模块，作为通信模块，用于将所述心电信号、所述脉搏波信号、所述加速度信号和所述心律失常诊断结果传输至移动终端；电源管理模块，包括充电、供电和开关电路，用于为所述系统提供电源供应。本发明提供的系统稳定、便携、功耗低，对心血管疾病的预防和物联网心电监测具有重要意义。
The present invention relates to the field of intelligent medical technology, and in particular to an arrhythmia diagnosis system based on multi-sensor fusion, the system comprising: an electrocardiogram signal module, comprising an electrocardiogram signal sensor, for collecting electrocardiogram signals; a pulse wave signal module, comprising a pulse wave sensor, for collecting pulse wave signals; a three-axis acceleration module, comprising an acceleration sensor, for collecting acceleration signals; a main control chip module, as a data processing module, for data reception and data analysis, to obtain arrhythmia diagnosis results; a Bluetooth module, as a communication module, for transmitting the electrocardiogram signal, the pulse wave signal, the acceleration signal and the arrhythmia diagnosis results to a mobile terminal; a power management module, comprising a charging, power supply and switching circuit, for providing power supply for the system. The system provided by the present invention is stable, portable, and has low power consumption, and is of great significance for the prevention of cardiovascular diseases and Internet of Things electrocardiogram monitoring.
Description
Technical Field
The invention relates to the technical field of intelligent medical treatment, in particular to an arrhythmia diagnosis system based on multi-sensor fusion.
Background
Heart disease is a disease which has great threat to human health, and an electrocardiogram is taken as an examination means for recording the electrical activity of the heart, so that the whole process of the heart can be reflected; clinically, 12-lead systems are mostly adopted, but the traditional 12-lead equipment has portability problems; furthermore, the wearable electrocardiograph equipment becomes a research hot spot, and with the rapid development of the wearable computing technology, great influence is generated on the health and medical field, and the wearable equipment is small, exquisite and portable, and is suitable for daily long-term monitoring; however, the wearable equipment still needs to be optimized in terms of precision and reliability, and a sensor technology needs to be improved to meet wider application requirements.
At present, research on electrocardiosignal filtering algorithms has achieved a certain result, however, in the use process of wearable electrocardiosignal monitoring equipment, baseline drift and motion artifact interference are often generated due to human body activities, particularly, the electrocardiosignal frequency in motion is similar to the interference frequency and is difficult to separate, a traditional self-adaptive filtering method is poor in performance when processing nonstationary noise and rapid change signals, and a signal processing algorithm needs to be improved.
The traditional electrocardiographic analysis relies on manual labeling of experts, is time-consuming and labor-consuming and is easily influenced by subjective factors, in recent years, an electrocardiographic characteristic can be automatically extracted by a deep learning-based method to realize efficient and accurate arrhythmia classification, but the complexity of a model is increased to bring challenges of calculation resources, therefore, the invention discloses an arrhythmia classification model, a dynamic rolling and multi-scale characteristic fusion strategy is introduced to reduce the complexity of the model, the arrhythmia is effectively classified, the requirement of calculation resources is reduced, a user can diagnose in real time, data is synchronized to a mobile terminal, the remote monitoring of professional medical staff and family members is facilitated, and a convenient, low-cost and comfortable technical scheme is provided for preventing heart health risks.
Disclosure of Invention
Aiming at the defects in the prior art and combining with the requirements of practical application, the invention provides an arrhythmia diagnosis system based on multi-sensor fusion, which comprises an electrocardiosignal module, a pulse wave signal module, a triaxial acceleration module, a main control chip module and a power management module, wherein the electrocardiosignal module comprises an electrocardiosignal sensor and is used for acquiring an electrocardiosignal of a subject, the pulse wave signal module comprises a pulse wave sensor and is used for acquiring a pulse wave signal of the subject, the triaxial acceleration module comprises an acceleration sensor and is used for acquiring an acceleration signal of the subject, the main control chip module is used as a data processing module of the system and is used for data receiving and data analysis so as to obtain an arrhythmia diagnosis result, the Bluetooth module is used as a communication module of the system and is used for transmitting the electrocardiosignal, the pulse wave signal, the acceleration signal and the arrhythmia diagnosis result to a mobile terminal, and the power management module comprises a charging, power supply and a switching circuit and is used for providing power supply for the system. According to the invention, by integrating the electrocardiosignal sensor, the pulse wave sensor and the triaxial acceleration sensor, the synchronous acquisition and fusion analysis of the multisource physiological signals are realized, the accuracy of arrhythmia diagnosis is obviously improved, the main control chip module carries out intelligent processing on the composite signals, interference is effectively filtered, the diagnosis result is more reliable, the Bluetooth module supports real-time data transmission to the mobile terminal, a user and a doctor can conveniently monitor the heart health state remotely, all-weather continuous monitoring is realized, the energy consumption is optimized by the power management module, the long-term effective operation of the system is ensured, portability and comfortableness are considered, and a high-efficiency and convenient solution is provided for preventing heart diseases and daily management.
Optionally, the main control chip module comprises an electrocardiosignal filtering module which is used as an input signal and used for preprocessing the input signal, and an arrhythmia diagnosis module which is used for diagnosing arrhythmia of the subject according to the input signal to obtain an arrhythmia diagnosis result. According to the invention, the original electrocardiosignal is preprocessed through the electrocardiosignal filtering module, noise such as motion artifact and myoelectric interference is effectively filtered, the signal-to-noise ratio is obviously improved, a foundation is laid for accurate diagnosis, the arrhythmia diagnosis module adopts an algorithm model to conduct intelligent analysis on the electrocardiosignal, the characteristics are automatically extracted and abnormal heart rhythm is identified through a deep learning model, the diagnosis accuracy is greatly improved, the real-time processing efficiency is ensured, the diagnosis reliability is improved, the system can still stably output professional diagnosis results in a complex environment, and high-efficiency and accurate technical support is provided for heart health monitoring.
Optionally, the electrocardiosignal filtering module comprises a filtering algorithm unit and an algorithm experiment unit, wherein the filtering algorithm unit is used for constructing an optimized filtering algorithm according to the self-adaptive mixed step factor and the noise power estimation mechanism, and the algorithm experiment unit is used for filtering the input signal according to the optimized filtering algorithm. The invention introduces a self-adaptive mixed step factor and noise power estimation mechanism through a filtering algorithm unit, so that the filtering algorithm can be dynamically matched with signal characteristics, stronger adaptability and robustness are shown in a complex noise environment, an algorithm experiment unit effectively filters noise such as motion artifact, myoelectric interference and the like through repeated verification and optimizes the filtering algorithm, the signal-to-noise ratio of an electrocardiosignal is obviously improved, the two units cooperate to ensure the real-time performance of the filtering process, greatly improve the signal quality, provide a high-fidelity data basis for subsequent arrhythmia diagnosis, ensure the filtering stability of the system in a complex scene and obviously enhance the diagnosis reliability.
Optionally, the construction of the optimized filtering algorithm according to the adaptive mixed step size factor and the noise power estimation mechanism comprises the steps of constructing a dynamic step size factor according to an error signal and the input signal based on the adaptive mixed step size factor, dynamically adjusting the dynamic step size factor according to the noise power estimation mechanism, and constructing the optimized filtering algorithm by combining a mixing strategy. According to the invention, an optimized filtering algorithm is constructed through the dynamic step factor, parameters are adjusted in real time according to error signals and input signals, the adaptability of the algorithm to signal change is obviously enhanced, the step factor is dynamically optimized by the noise power estimation mechanism, the filtering strength is finely adjusted according to the real-time noise level, so that the algorithm maintains high-efficiency denoising capability in a complex noise environment, the fidelity of the filtering signals is ensured, the processing efficiency is improved by combining a mixing strategy, the filtering process has flexibility and accuracy, diversified interference is effectively treated, high-quality electrocardiosignals are provided for subsequent diagnosis, and the accuracy and system stability of arrhythmia detection are greatly improved.
Optionally, the constructing a dynamic step factor based on the adaptive mixing step factor according to an error signal and the input signal includes:
Wherein, the As a dynamic step-size factor,As a result of the initial step size factor,As the energy of the error signal,In order to be able to input the energy of the signal,For the purpose of the noise power estimation,Is a small positive number. The invention enables the filtering algorithm to respond to the change of the signal and the noise in real time through the dynamic step factor, enables the algorithm to automatically reduce the step length to avoid overshoot when the noise is high, increases the step length to accelerate convergence when the noise is low, obviously improves the stability and the precision of filtering, effectively balances the noise suppression and the signal fidelity, and enables the algorithm to show stronger self-adaptive capacity in complex physiological signal environments.
Optionally, the filtering the input signal according to the optimized filtering algorithm includes obtaining the electrocardiosignal, the pulse wave signal and the acceleration signal of the subject in a jogging state, using the pulse wave signal and the acceleration signal as reference input signals, passing the reference input signals through a disturbance cancellation system, building write value coefficient iterative expression based on an adaptive algorithm to adjust a weight coefficient of the reference input signals, and performing loop iteration on the input signals based on the weight coefficient iterative expression by using the optimized filtering algorithm to realize filtering of the input signals. The invention dynamically adjusts the signal weight coefficient by collecting the multisource physiological signal in the jogging state and taking the pulse wave and the acceleration signal as reference inputs and utilizing the interference cancellation system to combine the self-adaptive algorithm, effectively eliminates the motion interference, optimizes the filtering algorithm to carry out loop iteration based on the weight coefficient, further purifies the electrocardiosignal, remarkably improves the signal-to-noise ratio, not only retains the key characteristics of the electrocardiosignal, but also stably outputs high-quality signals in complex motion scenes, provides reliable basis for subsequent arrhythmia diagnosis, and ensures that the monitoring system has robustness and practicability in practical application.
Optionally, the step of passing the reference input signal through a destructive interference system and constructing write value coefficient iterative expression based on an adaptive algorithm includes:
Wherein, the Is the firstThe weight coefficient at the time of the iteration,Is the firstThe weight coefficient at the time of the iteration,As a dynamic step-size factor,Is the firstThe input signal at the time of the iteration,Is the firstThe error signal at the time of the iteration,In order to be able to input the energy of the signal,Is a small positive number. The invention enables the filtering algorithm to respond to the change of the signal and the error in real time through the dynamic adjustment of the weight coefficient, enables the algorithm to automatically adjust the weight to strengthen the filtering when the error is increased, keeps the weight to keep details when the signal is stable, obviously improves the precision and stability of the self-adaptive filtering, effectively balances the noise suppression and the signal fidelity, enables the algorithm to show stronger anti-interference capability under complex motion scenes, and improves the overall performance of the system.
Optionally, the arrhythmia diagnosis module comprises a model construction unit and a model application unit, wherein the model construction unit is used for constructing an arrhythmia classification model based on a dynamic convolution and multi-scale feature fusion strategy, and the model application unit is used for training the arrhythmia classification model according to whole sample data and obtaining arrhythmia classification results. The model construction unit adopts a dynamic convolution and multi-scale feature fusion strategy, so that the classification model can adaptively adjust convolution kernel parameters, transient features in electrocardiosignals can be accurately captured, physiological information with different scales is fused, the accuracy of arrhythmia classification is remarkably improved, and the model application unit ensures that the model covers wide physiological scenes and enhances generalization capability through integral sample data training. The diagnosis system can not only identify complex arrhythmia modes, but also adapt to individual differences, keep high performance in diversified clinical data, and provide high-efficiency and accurate intelligent assistance for heart disease screening.
The method comprises the steps of replacing a one-dimensional convolution layer with a dynamic one-dimensional convolution layer, dynamically adjusting convolution kernels according to an input signal and combining the dynamic one-dimensional convolution layer, introducing the multi-dimensional convolution layer into a convolution neural network, extracting multi-scale features according to the multi-scale convolution layers based on the convolution kernels with different sizes, carrying out feature fusion on the multi-scale features based on the multi-scale feature fusion strategy to construct fusion features, and taking the fusion features as feature input of a converter module to construct the arrhythmia classification model. According to the invention, the dynamic one-dimensional convolution layer is adopted, the convolution kernel parameters are adjusted in real time according to the characteristics of the input signals, the adaptability of the model to signal changes is remarkably enhanced, the model can capture local details and global features at the same time through the design of the multi-scale convolution layer, multi-scale information is extracted through convolution kernels with different sizes, the richness of feature expression is improved, the feature fusion strategy further integrates the multi-scale information, and a more characteristic fusion feature is formed, so that the model shows higher classification precision in complex arrhythmia pattern recognition.
Optionally, training the arrhythmia classification model according to the whole sample data and obtaining an arrhythmia classification result, wherein the training of the arrhythmia classification model comprises the steps of dividing the whole sample data to construct a training set and a testing set, training the arrhythmia model, and obtaining the arrhythmia classification result according to the arrhythmia classification model after training to serve as the arrhythmia diagnosis result. According to the invention, the whole sample data are divided into the training set and the test set, so that the risk of overfitting is effectively avoided, the model is ensured to have good generalization capability in diversified data, the training set is utilized to fully optimize the arrhythmia classification model, so that model parameters are more fit with actual data distribution, the classification precision is remarkably improved, the performance of the model is objectively evaluated based on independent verification of the test set, the reliability of a diagnosis result is ensured, the recognition capability of the model to different arrhythmia types is improved, the practicability of the system in clinical application is enhanced, and high-efficiency and accurate technical support is provided for arrhythmia diagnosis.
Drawings
FIG. 1 is a hardware block diagram of an arrhythmia diagnostic system based on multi-sensor fusion according to an embodiment of the invention;
FIG. 2 is a block diagram of an algorithm implementation of an improved NLMS algorithm in accordance with an embodiment of the present invention;
FIG. 3 is a signal waveform diagram of a subject in jogging status according to an embodiment of the present invention;
FIG. 4 is a graph showing the mean square error of a filtering algorithm according to an embodiment of the present invention;
FIG. 5 is a comparison chart of the filtering waveforms of the filtering algorithm according to the embodiment of the present invention;
FIG. 6 is a model block diagram of an arrhythmia classification model according to an embodiment of the invention;
FIG. 7 is a graph of the change in the loss function of an arrhythmia model during training according to an embodiment of the invention;
FIG. 8 is a graph of the accuracy rate change of an arrhythmia model during training according to an embodiment of the invention;
FIG. 9 is a confusion matrix diagram of an arrhythmia classification model according to an embodiment of the invention;
Fig. 10 is a mobile terminal information display diagram according to an embodiment of the present invention;
Detailed Description
Specific embodiments of the invention will be described in detail below, it being noted that the embodiments described herein are for illustration only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that these specific details need not be employed to practice the present invention. In other instances, well-known circuits, software, or methods have not been described in detail in order not to obscure the invention.
Reference throughout this specification to "one embodiment," "an embodiment," "one example," or "an example" means that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the present invention. Thus, the appearances of the phrases "in one embodiment," "in an embodiment," "one example," or "an example" in various places throughout this specification are not necessarily all referring to the same embodiment or example. Furthermore, the particular features, structures, or characteristics may be combined in any suitable combination and/or sub-combination in one or more embodiments or examples. Moreover, those of ordinary skill in the art will appreciate that the illustrations provided herein are for illustrative purposes and that the illustrations are not necessarily drawn to scale.
Referring to fig. 1, according to the design requirements of a portable system, a set of hardware circuit system with excellent performance is constructed as a carrier for signal processing by combining signal acquisition, algorithm and wireless transmission, one embodiment of the invention provides an arrhythmia diagnosis system based on multi-sensor fusion, which comprises an electrocardiosignal module, a pulse wave signal module, a triaxial acceleration module, a master control chip module and a Bluetooth module, wherein the electrocardiosignal module comprises an electrocardiosignal sensor and is used for acquiring electrocardiosignal signals of a subject, the pulse wave signal module comprises a pulse wave sensor and is used for acquiring pulse wave signals of the subject, the triaxial acceleration module comprises an acceleration sensor and is used for acquiring acceleration signals of the subject, the master control chip module is used as a data processing module of the system and is used for data receiving and data analyzing so as to obtain arrhythmia diagnosis results, the Bluetooth module is used as a communication module of the system and is used for transmitting the electrocardiosignal, the pulse wave signals, the acceleration signals and the arrhythmia diagnosis results to a mobile terminal, and the power management module comprises a charging, a power supply and a switching circuit and is used for providing power for the system.
In the embodiment, an electrocardiosignal module is connected with a main control chip module through a Serial Peripheral Interface (SPI) by using a high-precision ADS1292R chip of Texas instruments, a pulse wave signal module is connected with the main control chip module through an integrated circuit bus (I2C), a triaxial acceleration module is connected with the main control chip module through an I2C by using an ADXL345 chip, the main control chip module is an ultralow-power-consumption STM32L452 series chip of an intentional semiconductor, a Bluetooth module is connected with the main control chip module through a V4.0BLE CC2540F256 chip by using a universal asynchronous receiver/transmitter (UART), a power management module comprises a charging circuit, a power supply and switching circuit, the charging circuit selects a BQ21040, the power supply voltage stabilizing circuit selects a TLV70033DCK voltage stabilizing chip, and the switching circuit realizes switching of a charging state and a power supply state of the system by using a PMOS tube, so that the power consumption of the system is reduced to increase the endurance time of an electrocardio detection device.
In the arrhythmia diagnosis system based on multi-sensor fusion provided by the invention, the main control chip module comprises:
A1, an electrocardiosignal filtering module, wherein the electrocardiosignal is used as an input signal, and the electrocardiosignal filtering module is used for preprocessing the input signal.
The electrocardiosignal filtering module comprises:
And A11, a filtering algorithm unit, wherein the filtering algorithm unit is used for constructing an optimized filtering algorithm according to the self-adaptive mixed step factor and the noise power estimation mechanism.
In the prior art, research on an electrocardiosignal filtering algorithm has achieved a certain result, however, in the use process of the wearable electrocardiosignal monitoring equipment, baseline drift and motion artifact interference are often generated due to human body activities, and particularly, the electrocardiosignal frequency in motion is similar to the interference frequency and is difficult to separate. Traditional self-adaptive filtering methods, such as LMS algorithm and NLMS algorithm, are widely applied due to simple calculation and good performance, but perform poorly when processing non-stationary noise and fast-varying signals, and traditional NLMS algorithm uses fixed step factors, resulting in limited convergence speed and stability of the filter under different noise environments.
In this embodiment, an improved NLMS algorithm is provided, and an adaptive mixing step factor and a noise power estimation mechanism are introduced, so that robustness and adaptability of the algorithm in a complex noise environment are improved by combining a dynamic step factor, noise power estimation and a mixing strategy.
Specifically, the adaptive blend step size factor is dynamically adjusted based on historical information of the error signal and the energy of the input signal. The current step length is determined by calculating the square of the current error and the square of the historical error in real time, so that the algorithm can quickly respond to the noise environment change.
Furthermore, in order to improve the robustness of the algorithm, a noise power estimation mechanism is introduced, wherein the noise power estimation mechanism reduces the step length when the noise is high so as to avoid excessive adjustment, and increases the step length when the noise is low so as to accelerate the convergence speed, thereby ensuring the stability and the efficiency of the filter.
The dynamic step factor is calculated based on the energy of the current error signal and the energy of the input signal, satisfying the following relationship:
Wherein, the As a dynamic step-size factor,As a result of the initial step size factor,As the energy of the error signal,In order to be able to input the energy of the signal,For the purpose of the noise power estimation,Is a small positive number.
The energy of the input signal satisfies the following relationship:
Wherein, the In order to be able to input the energy of the signal,Is the firstAn input signal at a time of iteration.
It should be noted that the number of the substrates,Is a small positive number used to avoid zero.
In this embodiment, a hybrid strategy is provided in combination with the advantages of the dynamic step size factor and the noise power estimation, in each iteration, the direction and the amplitude of the weight update are adaptively adjusted according to the energy of the current error signal and the input signal, the dynamic step size factor is dynamically adjusted according to the noise power estimation mechanism, and an optimized filtering algorithm is constructed in combination with the hybrid strategy, so that the adaptive step size factor is used for improving the NLMS algorithm and the operand is not greatly increased.
Referring to fig. 2, a block diagram of an algorithm implementation for improving the NLMS algorithm is illustrated, where a pulse wave signal and an acceleration signal are used as reference input signals, and an electrocardiograph signal is used as an original input signal.
Specifically, the filter output satisfies the following relationship:
Wherein, the In order for the output of the filter to be present,For the output of the sub-filter,Is the pulse wave signal at the firstThe weight coefficient at the time of the iteration,In order to input a signal to the device,Is the acceleration signal at the firstThe weight coefficient at the time of the iteration,As an error signal, the signal is a signal,To output a desired signal.
And A12, an algorithm experiment unit, wherein the algorithm experiment unit is used for filtering the input signal according to the optimized filtering algorithm.
In this example, to verify the effectiveness of the modified NLMS algorithm, test and comparative analysis were performed experimentally.
The algorithm experiment unit executes the following steps:
S1, acquiring the electrocardiosignals, the pulse wave signals and the acceleration signals of the subject in a jogging state, wherein the pulse wave signals and the acceleration signals are used as reference input signals.
Referring to fig. 3, a signal waveform diagram of a subject in jogging state is illustrated, including an electrocardiographic signal waveform diagram in fig. 3 (a), a pulse wave signal waveform diagram in fig. 3 (b) and an acceleration signal waveform diagram in fig. 3 (c), and the electrocardiographic signal sensor, the pulse wave sensor and the acceleration sensor are used for collecting signals respectively, wherein the sampling frequency is 500Hz.
S2, the reference input signal passes through a disturbance cancellation system, and a write value coefficient iterative expression is built based on a self-adaptive algorithm so as to adjust the weight coefficient of the reference input signal.
Specifically, based on a disturbance cancellation system, an iterative expression of a weight coefficient is constructed according to a dynamic step factor by combining an adaptive algorithm, and the following relation is satisfied:
Wherein, the Is the firstThe weight coefficient at the time of the iteration,Is the firstThe weight coefficient at the time of the iteration,As a dynamic step-size factor,Is the firstThe input signal at the time of the iteration,Is the firstThe error signal at the time of the iteration,In order to be able to input the energy of the signal,Is a small positive number.
Based on an improved NLMS algorithm, the weight coefficient of the pulse wave signal and the acceleration signal is adjusted by combining the weight coefficient iterative expression, so that the following relation is satisfied:
Wherein, the Is the pulse wave signal at the firstThe weight coefficient at the time of the iteration,Is the pulse wave signal at the firstThe weight coefficient at the time of the iteration,Is the firstThe dynamic step size factor at the time of the iteration,As an error signal, the signal is a signal,In order to input a signal to the device,The transpose is represented by the number,Is a small positive number, the number of which is,Is the acceleration signal at the firstThe weight coefficient at the time of the iteration,Is the acceleration signal at the firstWeight coefficient at the time of iteration.
In an alternative embodiment, the LMS algorithm, the NLMS algorithm, and the weight coefficient updating process calculation amount of the modified NLMS algorithm are compared to obtain the filtering effect of each algorithm, as shown in table 1:
TABLE 1
Further, referring to fig. 4, a mean square error comparison graph of a filtering algorithm is illustrated, which includes a LMS algorithm, an NLMS algorithm, and a mean square error learning curve for improving the NLMS algorithm.
By comprehensively analyzing the table 1 and the figure 4, experimental results show that compared with an NLMS algorithm, the improved NLMS algorithm shows higher robustness and adaptability in noise removal, the signal-to-noise ratio SNR of the improved NLMS algorithm reaches 18.853dB, the signal-to-noise ratio SNR of the improved NLMS algorithm is improved by 16% compared with the NLMS algorithm, and compared with the LMS algorithm, the improved NLMS algorithm disclosed by the invention is obviously superior to the LMS algorithm and the NLMS algorithm, and in non-stationary signals and fast-changing noise environments, the improved algorithm can converge to a steady state more quickly, and the denoising effect is good.
S3, carrying out loop iteration on the input signal based on the weight coefficient iteration expression by utilizing the optimized filtering algorithm so as to realize filtering of the input signal.
In this embodiment, python is used to design LMS algorithm, NLMS algorithm and modified NLMS algorithm respectively, and in jogging state, loop iteration is performed on the electrocardiosignal containing motion artifact interference and baseline drift noise by using weight coefficient iteration expression, and the signals are filtered by LMS algorithm, NLMS algorithm and modified NLMS algorithm respectively to obtain a filtered waveform.
Referring to fig. 5, a comparison diagram of the filtered waveforms of the filtering algorithm is illustrated, including the LMS algorithm, the NLMS algorithm, and the modified NLMS algorithm.
Further, the weight coefficient updates of the LMS algorithm, the NLMS algorithm, and the modified NLMS algorithm were compared and analyzed as shown in table 2:
TABLE 2
By comprehensively analyzing the combination of fig. 5 and table 2, it can be seen that the operation amount of the LMS algorithm is minimum, but the method is only suitable for environments with limited computing resources, and compared with the LMS algorithm, the NLMS algorithm adds additional multiplication and addition operations, so that the stability and convergence speed of the algorithm are improved, the improved NLMS algorithm further adds a small amount of multiplication and addition operations, the characteristic of dynamically adjusting step factors is obviously improved, and the performance of the algorithm is particularly enhanced under different noise environments.
A2, an arrhythmia diagnosis module is used for carrying out arrhythmia diagnosis on the subject according to the input signal to obtain an arrhythmia diagnosis result.
The arrhythmia diagnostic module comprises:
a21, a model building unit, which builds an arrhythmia classification model based on a dynamic convolution and a multi-scale feature fusion strategy;
In the prior art, traditional electrocardiographic analysis relies on manual labeling by an expert, which is time-consuming, laborious and subject to subjective factors. In recent years, the deep learning-based method can automatically extract the electrocardio features to realize efficient and accurate arrhythmia classification, but the increase of model complexity also brings about the challenge of computing resources.
In the embodiment, a dynamic convolution and multi-scale feature fusion strategy is introduced to construct an arrhythmia classification model so as to reduce model complexity, and a 4070ti display card and an i7-13700kf processor are adopted by a device computer.
Referring to fig. 6, a model structure diagram of an arrhythmia classification model is shown, an improved CNN-Transforme is obtained based on CNN-Transforme and is used as an arrhythmia classification model, firstly, an input layer is used for inputting an input signal, the input signal passes through two one-dimensional convolution layers and is subjected to batch normalization, secondly, the one-dimensional convolution layer is replaced by a dynamic one-dimensional convolution layer, the convolution kernel is dynamically adjusted by combining with the dynamic one-dimensional convolution layer, the quantity of parameters and the calculated quantity are reduced, then, a multi-scale convolution layer is introduced into a convolution neural network through a one-dimensional maximum pooling layer, multi-scale features are extracted according to the multi-scale convolution layer based on different sizes of convolution kernels, the characterization capability of the model is enhanced, and feature fusion is carried out on the multi-scale features based on a multi-scale feature fusion strategy, then, the fusion features are used as feature input of a converter module, the time sequence dependency relationship in the signal is captured, and finally, the multi-scale feature fusion model is output through a flattening layer, a discarding layer and two full-connection layers.
And A22, a model application unit, which is used for training the arrhythmia classification model according to the whole sample data and obtaining an arrhythmia classification result.
In this embodiment, the arrhythmia classification model is trained, and the whole sample data is divided, with 80% of the whole sample data as a training set and 20% of the whole sample data as a test set.
Referring to fig. 7, a graph of the change in the loss function of the arrhythmia model during training is illustrated, showing the change in the loss function of the arrhythmia model with increasing number of iterations, including a training loss curve and a test loss curve.
Referring to fig. 8, a graph of the change in accuracy of the arrhythmia model during training is shown, showing the change in accuracy of the arrhythmia model with increasing number of iterations, including training accuracy and testing accuracy.
By comprehensively analyzing with the combination of fig. 7 and 8, with the increase of the iteration times, the accuracy in the training process can be observed to continuously rise, the rising speed is slower and slower, the training process finally tends to be stable, the loss function in the training process changes inversely, the training process finally tends to be a stable value, the accuracy of the final arrhythmia model on the training set reaches 99.2%, and the accuracy on the verification set reaches 98%.
Further, an arrhythmia classification result is obtained based on the trained arrhythmia classification model and is used as an arrhythmia diagnosis result.
In an alternative embodiment, the test set is used to verify the performance of the arrhythmia classification model after the training process of the arrhythmia classification model is completed.
Referring to fig. 9, a confusion matrix diagram of an arrhythmia classification model is illustrated, which shows the arrhythmia classification model comparing N, L, R, A and V five electrocardiograph signal prediction labels with real labels, comparing arrhythmia classification results of the arrhythmia classification model with labels of MIT-BIH arrhythmia database data, using a confusion matrix form to represent that the real class is correctly classified as the sample number of the real class, and the rest is the sample number of the misclassified class.
Further, the accuracy, precision, recall and F1 values of the arrhythmia classification model on N, L, R, A and V electrocardiosignals are calculated according to the data in the confusion matrix, and evaluation index values are obtained, as shown in Table 3:
TABLE 3 Table 3
As shown in Table 3, the classification accuracy rate reaches more than 98.6%, the classification accuracy rate of each label reaches more than 95.3%, the recall rate of each label reaches more than 95%, and the F1 value of each label reaches more than 96.3%, so that the improved CNN-transducer arrhythmia classification model has excellent classification performance and good indexes, and basically meets the requirements of arrhythmia classification prediction.
Furthermore, in order to verify the advancement and accuracy of the arrhythmia classification model disclosed by the invention, the model improvability is better explored, and a comparison experiment is carried out on a DNN model, a CNN model and a CNN-transducer model respectively by adopting the same data set, so that the accuracy performance of the three models and the improved CNN-transducer arrhythmia classification model disclosed by the invention under the same data set is compared, and the accuracy comparison result is shown in table 4:
TABLE 4 Table 4
As can be seen from Table 4, the improved CNN-transducer arrhythmia classification model adopted by the invention has excellent classification performance, which shows that the arrhythmia classification model has better effect in electrocardiogram classification work.
In another alternative embodiment, the arrhythmia diagnosis system (multi-sensor diagnosis terminal) based on multi-sensor fusion provided by the invention is subjected to actual tests and result analysis, and in the actual tests, 10 groups of experimental data of 18-30 years old subjects are collected, wherein 7 groups of men, 3 groups of women and participants have no dyskinesia and abnormal heart rate.
Firstly, connecting a multi-sensor diagnosis terminal to a subject, wearing a pulse wave sensor by a left index finger, fixing an acceleration sensor on the waist of the human body because the waist position is close to the gravity center of the human body, and wearing an electrocardiosignal sensor connection electrode patch on the left chest, the right chest and the right abdomen respectively, synchronously acquiring electrocardiosignals, pulse wave signals and acceleration signals for 1 minute through wearing the pulse wave sensor, the electrocardiosignal sensor and the acceleration sensor, storing the electrocardiosignals, the pulse wave signals, the acceleration signals and the acceleration signals as CSV files, carrying out classification prediction on data by using a trained improved CNN-transducer arrhythmia classification model after denoising through multi-sensor information fusion, wherein the electrocardiosignal classification probability is shown in a table 5:
TABLE 5
As can be seen from Table 5, the improved CNN-transducer arrhythmia classification model provided by the invention has better classification prediction performance.
In order to display physical sign signals in real time, an Android platform APP is developed, electrocardiosignals, pulse wave signals and acceleration signals can be displayed through a Bluetooth module, electrocardiosignals are monitored in real time, and functions of signal display, data storage, historical data review and the like are achieved.
In summary, the arrhythmia diagnosis system based on multi-sensor fusion is provided, a wearable multi-sensor information fusion electrocardiograph monitoring system is designed under the background that the incidence rate of cardiovascular diseases continuously rises, a hardware acquisition system based on an STM32L452CEU6 chip is designed, electrocardiographic signals, pulse wave signals and acceleration signals are acquired by using ADS1292R and ADXL345 chips and transmitted to a mobile terminal through a Bluetooth module, an improved NLMS algorithm is provided, the self-adaptive step size factor and a noise power estimation mechanism are adopted, acceleration and pulse wave signals are fused for filtering, the signal-to-noise ratio is improved by 16% to achieve 18.853dB, a CNN-transducer arrhythmia classification model is built by the improved CNN-transducer algorithm, a dynamic rolling and multi-scale feature fusion strategy is introduced, a lightweight model is realized, the classification accuracy rate is more than 98.6%, and the test result shows that the system is stable, portable and low in power consumption and has important significance for preventing cardiovascular diseases and monitoring of Internet of things.
It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, but not to limit the technical solution of the present invention, and although the detailed description of the present invention is given with reference to the above embodiments, it should be understood by those skilled in the art that the technical solution described in the above embodiments may be modified or some or all technical features may be equivalently replaced, and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiments of the present invention, and all the modifications or substitutions are included in the scope of the claims and the specification of the present invention.
Claims (10)
1. An arrhythmia diagnostic system based on multisensor fusion, comprising:
The electrocardiosignal module comprises an electrocardiosignal sensor and is used for acquiring electrocardiosignals of a subject;
A pulse wave signal module including a pulse wave sensor for acquiring a pulse wave signal of the subject;
the triaxial acceleration module comprises an acceleration sensor and is used for acquiring acceleration signals of the subject;
the main control chip module is used as a data processing module of the system and used for data receiving and data analysis so as to obtain arrhythmia diagnosis results;
The Bluetooth module is used as a communication module of the system and is used for transmitting the electrocardiosignal, the pulse wave signal, the acceleration signal and the arrhythmia diagnosis result to a mobile terminal;
and the power management module comprises a charging, power supplying and switching circuit and is used for providing power supply for the system.
2. The multi-sensor fusion based arrhythmia diagnostic system of claim 1 wherein the main control chip module comprises:
an electrocardiosignal filtering module, wherein the electrocardiosignal is used as an input signal, the electrocardiosignal filtering module is used for preprocessing the input signal;
and the arrhythmia diagnosis module is used for carrying out arrhythmia diagnosis on the subject according to the input signal to obtain an arrhythmia diagnosis result.
3. The multi-sensor fusion based arrhythmia diagnostic system of claim 2 wherein the electrocardiosignal filtering module comprises:
The filtering algorithm unit is used for constructing an optimized filtering algorithm according to the self-adaptive mixed step factor and the noise power estimation mechanism;
and the algorithm experiment unit is used for filtering the input signal according to the optimized filtering algorithm.
4. The multi-sensor fusion based arrhythmia diagnostic system of claim 3 wherein the constructing an optimized filtering algorithm based on an adaptive hybrid step size factor and noise power estimation mechanism comprises:
based on the adaptive mixing step size factor, constructing a dynamic step size factor according to an error signal and the input signal;
And dynamically adjusting the dynamic step factor according to the noise power estimation mechanism, and constructing the optimized filtering algorithm by combining a mixing strategy.
5. The multiple sensor fusion based arrhythmia diagnostic system of claim 4 wherein the constructing a dynamic step factor from an error signal and the input signal based on the adaptive hybrid step factor comprises:
Wherein, the As a dynamic step-size factor,As a result of the initial step size factor,As the energy of the error signal,In order to be able to input the energy of the signal,For the purpose of the noise power estimation,Is a small positive number.
6. The multisensor fusion-based arrhythmia diagnostic system of claim 3, wherein filtering the input signal in accordance with the optimized filtering algorithm comprises:
acquiring the electrocardiosignal, the pulse wave signal and the acceleration signal of the subject in a jogging state, wherein the pulse wave signal and the acceleration signal are used as reference input signals;
The reference input signal passes through a disturbance cancellation system, and a write value coefficient iterative expression is built based on a self-adaptive algorithm so as to adjust the weight coefficient of the reference input signal;
And carrying out loop iteration on the input signal based on the weight coefficient iteration expression by utilizing the optimized filtering algorithm so as to realize the filtering of the input signal.
7. The multiple sensor fusion-based arrhythmia diagnostic system of claim 6 wherein said passing the reference input signal through a destructive interference system and constructing write value coefficient iterative expressions based on an adaptive algorithm comprises:
Wherein, the Is the firstThe weight coefficient at the time of the iteration,Is the firstThe weight coefficient at the time of the iteration,As a dynamic step-size factor,Is the firstThe input signal at the time of the iteration,Is the firstThe error signal at the time of the iteration,In order to be able to input the energy of the signal,Is a small positive number.
8. The multi-sensor fusion based arrhythmia diagnostic system of claim 2 wherein the arrhythmia diagnostic module comprises:
The model building unit is used for building an arrhythmia classification model based on the dynamic convolution and the multi-scale feature fusion strategy;
The model application unit is used for training the arrhythmia classification model according to the whole sample data and obtaining arrhythmia classification results.
9. The multisensor fusion-based arrhythmia diagnostic system of claim 8, wherein the constructing an arrhythmia classification model based on dynamic convolution and multiscale feature fusion strategy comprises:
replacing the one-dimensional convolution layer with a dynamic one-dimensional convolution layer, and dynamically adjusting a convolution kernel by combining the dynamic one-dimensional convolution layer according to the input signal;
introducing a multi-scale convolution layer into a convolution neural network, and extracting multi-scale features according to the multi-scale convolution layer based on the convolution kernels with different sizes;
performing feature fusion on the multi-scale features based on the multi-scale feature fusion strategy to construct fusion features;
And taking the fusion characteristic as the characteristic input of a converter module to construct the arrhythmia classification model.
10. The multi-sensor fusion based arrhythmia diagnostic system of claim 8 wherein the training the arrhythmia classification model based on global sample data and obtaining arrhythmia classification results comprises:
dividing the whole sample data to construct a training set and a testing set, and training the arrhythmia model;
And obtaining the arrhythmia classification result according to the trained arrhythmia classification model, and taking the arrhythmia classification result as the arrhythmia diagnosis result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510421947.8A CN120203550B (en) | 2025-04-07 | 2025-04-07 | An arrhythmia diagnosis system based on multi-sensor fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510421947.8A CN120203550B (en) | 2025-04-07 | 2025-04-07 | An arrhythmia diagnosis system based on multi-sensor fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN120203550A true CN120203550A (en) | 2025-06-27 |
| CN120203550B CN120203550B (en) | 2025-08-22 |
Family
ID=96114670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510421947.8A Active CN120203550B (en) | 2025-04-07 | 2025-04-07 | An arrhythmia diagnosis system based on multi-sensor fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120203550B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120802015A (en) * | 2025-09-08 | 2025-10-17 | 北京智慧大王科技有限公司 | Circuit break monitoring device and method for electrical safety protection |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050101889A1 (en) * | 2003-11-06 | 2005-05-12 | Freeman Gary A. | Using chest velocity to process physiological signals to remove chest compression artifacts |
| US20100113888A1 (en) * | 2008-10-30 | 2010-05-06 | Medtronic,Inc. | Heart failure decompensation determination |
| CN117369418A (en) * | 2023-11-17 | 2024-01-09 | 北京信息科技大学 | CNC machine tool feed system fault diagnosis method, system, storage media and equipment |
| CN117547230A (en) * | 2023-11-07 | 2024-02-13 | 北京工业大学 | Pulse wave conduction velocity acquisition method based on mobile phone front and rear cameras |
| CN117942056A (en) * | 2023-12-18 | 2024-04-30 | 复旦大学 | Arrhythmia multi-classification detection method and system based on pulse waves |
| CN117958774A (en) * | 2024-01-10 | 2024-05-03 | 国网江苏省电力有限公司无锡供电分公司 | A method and system for dynamic vital signs detection using photoelectric pulse wave signals |
| KR20240165711A (en) * | 2023-05-16 | 2024-11-25 | 주식회사 딥메트릭스 | Method for diagnosing arrhythmias using artificial intelligence |
| CN119701210A (en) * | 2025-02-26 | 2025-03-28 | 西安交通大学医学院第一附属医院 | Electronic collecting, storing and managing system for pacemaker program control data |
-
2025
- 2025-04-07 CN CN202510421947.8A patent/CN120203550B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050101889A1 (en) * | 2003-11-06 | 2005-05-12 | Freeman Gary A. | Using chest velocity to process physiological signals to remove chest compression artifacts |
| US20100113888A1 (en) * | 2008-10-30 | 2010-05-06 | Medtronic,Inc. | Heart failure decompensation determination |
| KR20240165711A (en) * | 2023-05-16 | 2024-11-25 | 주식회사 딥메트릭스 | Method for diagnosing arrhythmias using artificial intelligence |
| CN117547230A (en) * | 2023-11-07 | 2024-02-13 | 北京工业大学 | Pulse wave conduction velocity acquisition method based on mobile phone front and rear cameras |
| CN117369418A (en) * | 2023-11-17 | 2024-01-09 | 北京信息科技大学 | CNC machine tool feed system fault diagnosis method, system, storage media and equipment |
| CN117942056A (en) * | 2023-12-18 | 2024-04-30 | 复旦大学 | Arrhythmia multi-classification detection method and system based on pulse waves |
| CN117958774A (en) * | 2024-01-10 | 2024-05-03 | 国网江苏省电力有限公司无锡供电分公司 | A method and system for dynamic vital signs detection using photoelectric pulse wave signals |
| CN119701210A (en) * | 2025-02-26 | 2025-03-28 | 西安交通大学医学院第一附属医院 | Electronic collecting, storing and managing system for pacemaker program control data |
Non-Patent Citations (2)
| Title |
|---|
| 史健婷;黄剑华;张英涛;唐降龙;: "基于小波和自适应滤波的ECG基线漂移校正", 计算机工程, no. 11, 15 November 2013 (2013-11-15) * |
| 汤丽平;张健;刘剑;郭兴明;: "基于支持向量机的心律失常诊断研究", 激光杂志, no. 04, 15 August 2012 (2012-08-15) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120802015A (en) * | 2025-09-08 | 2025-10-17 | 北京智慧大王科技有限公司 | Circuit break monitoring device and method for electrical safety protection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN120203550B (en) | 2025-08-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104173043A (en) | A ECG data analysis method suitable for mobile platforms | |
| CN116687409B (en) | Emotion recognition method and system based on digital twin and deep learning | |
| CN120203550B (en) | An arrhythmia diagnosis system based on multi-sensor fusion | |
| CN105796091B (en) | A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise | |
| CN115251845B (en) | Sleep monitoring method for processing brain wave signals based on TB-TF-BiGRU model | |
| CN119046724A (en) | Method and system for detecting single-lead electrocardiogram arrhythmia based on self-adaptive attention | |
| Ngoc-Thang et al. | A dynamic reconfigurable wearable device to acquire high quality PPG signal and robust heart rate estimate based on deep learning algorithm for smart healthcare system | |
| Yuan et al. | Multivariate sleep stage classification using hybrid self-attentive deep learning networks | |
| Torti et al. | Custom FPGA processing for real-time fetal ECG extraction and identification | |
| Chen et al. | Multitask learning-based quality assessment and denoising of electrocardiogram signals | |
| Fu et al. | Finger gesture recognition using sensing and classification of surface electromyography signals with high-precision wireless surface electromyography sensors | |
| Rahman et al. | A smart wearable for real-time cardiac disease detection using beat-by-beat ECG signal analysis with an edge computing AI classifier | |
| Sá et al. | Heart disease detection architecture for lead i off-the-person ecg monitoring devices | |
| Wu et al. | A wireless surface EMG acquisition and gesture recognition system | |
| Vivekanand et al. | Tiny ML-based non-invasive approach of cardiac monitoring | |
| Kan et al. | A Novel PSO‐Based Optimized Lightweight Convolution Neural Network for Movements Recognizing from Multichannel Surface Electromyogram | |
| Hui et al. | Wearable devices acquired ECG signals detection method using 1D convolutional neural network | |
| Xie et al. | Design and implementation of an ARM-based AI module for ectopic beat classification using custom and structural pruned lightweight CNN | |
| Lin et al. | Portable ECG and PCG wireless acquisition system and multiscale CNN feature fusion Bi-LSTM network for coronary artery disease diagnosis | |
| Kumar et al. | Long short term memory (lstm)-based cuffless continuous blood pressure monitoring | |
| Liu et al. | Power and area efficient wavelet-based on-chip ECG processor for WBAN | |
| CN115886790A (en) | A method, device and equipment for spine posture detection based on sEMG and SpO2 | |
| Saadi et al. | Low-cost ecg monitoring system with classification using deep learning | |
| Rodríguez-Jorge et al. | Detection of the QRS Complexity in Real Time with Bluetooth Communication | |
| Sharma et al. | fECG extraction using 1D convolution neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |