CN116327181A - A comprehensive evaluation method, device and electronic equipment for real-time cardiac non-inductive monitoring - Google Patents

Abstract

Translated fromChinese

本发明公开了一种心脏实时无感监测综合评价方法，属于心脏监测领域，通过体征采集、对原始体动数据进行处理、搭建非接触心电合成模型、合成非接触的心电图数据以及对心脏进行评估等步骤，根据心脏机械运动信号和心脏电信号属同源不同表征，合成了实时、高精度的非接触心电图；基于所得非接触心电图和采集心冲击图，对心脏进行了多信号监测的基础上，进行综合评估和心血管疾病综合症的预测，基于该方法的合成心电图能够有效地保留心脏电生理功能信息、生物差异性以及病理信息。本发明还涉及实施上述心脏实时无感监测综合评价方法的装置以及电子设备。

The invention discloses a comprehensive evaluation method for real-time non-inductive monitoring of the heart, which belongs to the field of heart monitoring. It collects signs, processes original body movement data, builds a non-contact electrocardiographic synthesis model, synthesizes non-contact electrocardiogram data, and conducts heart monitoring. Evaluation and other steps, according to the homology and different characteristics of the cardiac mechanical motion signal and the cardiac electrical signal, a real-time, high-precision non-contact electrocardiogram was synthesized; based on the obtained non-contact electrocardiogram and collection of shockcardiogram, the basis for multi-signal monitoring of the heart On the basis of comprehensive evaluation and prediction of cardiovascular disease syndrome, the synthetic ECG based on this method can effectively retain the cardiac electrophysiological function information, biological differences and pathological information. The present invention also relates to a device and electronic equipment for implementing the above-mentioned comprehensive evaluation method for real-time non-inductive monitoring of the heart.

Description

Translated fromChinese

一种心脏实时无感监测综合评价方法、装置及电子设备A comprehensive evaluation method, device and electronic equipment for real-time cardiac non-inductive monitoring

技术领域technical field

本发明涉及心率监测领域，尤其是涉及心脏实时无感监测综合评价方法、系统以及装置。The invention relates to the field of heart rate monitoring, in particular to a comprehensive evaluation method, system and device for heart real-time non-inductive monitoring.

背景技术Background technique

心血管疾病的临床诊断前需要对心脏电生理、泵生理分别进行检测。电生理功能检测依靠心电图（ECG）评估心脏的电活动，泵生理功能检测技术包括超声心动图(Ultrasound cardiogram, UCG)、核磁共振（Magnetic resonance imaging, MRI）以评估心脏的血流动力学情况、心脏内部结构等信息。Cardiac electrophysiology and pump physiology need to be tested separately before the clinical diagnosis of cardiovascular disease. Electrophysiological function testing relies on electrocardiogram (ECG) to evaluate the electrical activity of the heart, and pump physiological function testing techniques include echocardiography (Ultrasound cardiogram, UCG) and magnetic resonance imaging (MRI) to evaluate the hemodynamic condition of the heart, Information about the internal structure of the heart.

尽管临床上两种生理功能检测技术互为补充，但是现有涉及心脏评估的移动医疗设备在非临床的日常家居场景中存在很多限制：Although the two physiological function detection technologies complement each other clinically, the existing mobile medical devices involving cardiac assessment have many limitations in non-clinical daily home scenarios:

一方面，传统心电图监测方法需要将传统接触式电极粘贴或夹紧在皮肤表面，通常情况下，为了得到良好的电生理信息,还需要在皮肤表面涂抹电解液，如专利CN210990273U所示，该心电图仪仍然需要佩戴电极片的采集方式不适用于因烧伤等原因造成体表皮肤破损的患者，同时基于电极片的佩戴方式也不适用于日常生活的长期使用。On the one hand, the traditional electrocardiogram monitoring method needs to paste or clamp the traditional contact electrode on the skin surface. Usually, in order to obtain good electrophysiological information, it is also necessary to apply electrolyte on the skin surface, as shown in the patent CN210990273U, the electrocardiogram The acquisition method that still needs to wear electrode pads is not suitable for patients with skin damage on the body surface due to burns and other reasons, and the wearing method based on electrode pads is not suitable for long-term use in daily life.

另一方面，近年市面上出现很多智能穿戴设备，比如，手环、手表、运动装备等设备通过采集腕部、胸前的心脏机械运动信号，对心率、呼吸率等周期性事件进行评估，但是仍无法代替作为心血管监测金标准的ECG，并且没有涉及对心脏多生理功能的综合评估。On the other hand, many smart wearable devices have appeared on the market in recent years. For example, wristbands, watches, sports equipment and other devices collect mechanical movement signals of the heart on the wrist and chest to evaluate periodic events such as heart rate and respiration rate. It still cannot replace ECG, which is the gold standard of cardiovascular monitoring, and it does not involve comprehensive assessment of multiple physiological functions of the heart.

因此，开发一种使用非接触式设备和深度学习的心脏监测以及综合评估方法是非常必要的。Therefore, it is highly necessary to develop a method for cardiac monitoring and comprehensive evaluation using non-contact devices and deep learning.

发明内容Contents of the invention

为了克服现有技术的不足，本发明的目的之一在于提供一种心脏实时无感监测综合评价方法，能够在不和皮肤直接接触的情况下捕捉到心脏的实时心电图数据。In order to overcome the deficiencies of the prior art, one of the objectives of the present invention is to provide a comprehensive evaluation method for real-time non-sensing monitoring of the heart, which can capture the real-time ECG data of the heart without direct contact with the skin.

为了克服现有技术的不足，本发明的目的之二在于提供一种心脏实时无感监测综合评价系统，能够在不和皮肤直接接触的情况下同时对心脏的电活动（包括不限于实时心电图）、电机械活动间隔（包括不限于射血前期）、血流动力学（包括不限于动态血压）进行综合评估。In order to overcome the deficiencies of the prior art, the second purpose of the present invention is to provide a real-time non-inductive monitoring comprehensive evaluation system for the heart, which can simultaneously monitor the electrical activity of the heart (including but not limited to real-time electrocardiogram) without direct contact with the skin. , electromechanical activity interval (including but not limited to pre-ejection period), hemodynamics (including but not limited to ambulatory blood pressure) for comprehensive evaluation.

为了克服现有技术的不足，本发明的目的之三在于提供一种心脏实时无感监测综合评价装置，能够在不和皮肤直接接触的情况下获取心脏数据，不影响人活动，实现房颤疾病的监测，并且对其血压等情况的监测可以实现并发血栓的风险。In order to overcome the deficiencies of the prior art, the third object of the present invention is to provide a real-time non-inductive monitoring and comprehensive evaluation device for the heart, which can obtain heart data without direct contact with the skin, without affecting people's activities, and achieve atrial fibrillation. The monitoring of blood pressure and other conditions can realize the risk of concurrent thrombosis.

本发明的目的之一采用如下技术方案实现：One of purpose of the present invention adopts following technical scheme to realize:

一种心脏实时无感监测综合评价方法，包括以下步骤：A comprehensive evaluation method for heart real-time sensorless monitoring, comprising the following steps:

体征采集：将柔性心脏带铺设在待监测人体胸部下方，利用柔性心脏带由于受到压力后产生微形变的特性，采集待监测人体的原始体动数据，并将原始体动数据无线传输至处理器；Sign collection: Lay the flexible heart belt under the chest of the human body to be monitored, and use the characteristics of the flexible heart belt to produce micro-deformation under pressure to collect the original body motion data of the human body to be monitored, and wirelessly transmit the original body motion data to the processor ;

对原始体动数据进行处理：处理器对原始体动数据中的心冲击信号BCG进行频谱分析，采用变分模态分解技术，提取BCG分量；Process the original body motion data: the processor performs spectrum analysis on the cardiac shock signal BCG in the original body motion data, and uses variational mode decomposition technology to extract the BCG component;

搭建非接触心电合成模型：对健康心脏的心电信号ECG进行心拍分割，并对分割后信号进行质量评估，采用质量好的分割后心电信号搭建非接触心电合成模型；Build a non-contact ECG synthesis model: Segment the heart rate of the ECG signal ECG of a healthy heart, and evaluate the quality of the segmented signal, and build a non-contact ECG synthesis model by using the segmented ECG signal with good quality;

合成非接触的心电图数据：将BCG分量输入至非接触心电合成模型，使用深度学习算法学习电-机械运动之间的非线性映射关系，合成非接触的心电图数据；Synthesize non-contact ECG data: input BCG components into the non-contact ECG synthesis model, use deep learning algorithm to learn the nonlinear mapping relationship between electrical-mechanical motion, and synthesize non-contact ECG data;

对心脏进行评估：基于非接触的心电图数据和实际测量的心冲击信号BCG，对人体从电生理、泵生理角度进行评估。Evaluate the heart: Based on the non-contact ECG data and the actual measured cardiac shock signal BCG, the human body is evaluated from the perspective of electrophysiology and pump physiology.

进一步地，在所述体征采集步骤中，所述柔性心脏带包括压电传感器，所述压电传感器由聚偏氟乙烯制成。Further, in the step of collecting signs, the flexible heart belt includes a piezoelectric sensor, and the piezoelectric sensor is made of polyvinylidene fluoride.

进一步地，在所述体征采集步骤中，采集待监测人体卧床时3种状态的原始体动数据，3种状态分别为：静息状态，待监测人体被要求统一以双手自然垂在身侧、保持均匀呼吸的姿势；运动后状态，待监测人体被要求进行若干次原地下蹲运动后卧床；睡眠状态，待监测人体处于睡眠中。Further, in the step of collecting signs, the original body movement data of three states of the human body to be monitored are collected when lying in bed. The three states are respectively: the resting state, and the human body to be monitored is required to hang both hands naturally by the side, Keep the posture of breathing evenly; in the post-exercise state, the human body to be monitored is required to lie in bed after performing several squat exercises; in the sleeping state, the human body to be monitored is in sleep.

进一步地，在对原始体动数据进行处理步骤中，采用变分模态分解技术，提取BCG分量具体为：Further, in the processing step of the original body motion data, the variational mode decomposition technique is used to extract the BCG component as follows:

变分模态分解公式为：The variational mode decomposition formula is:

（1）

(1)

（2）

(2)

其中

,表示不同模态；/>

，表示不同模态中心频率；/>

为狄拉克函数，K为需要分解的模态个数，e为自然底数，t为时间，j 表示复数，/>

表示原信号频率；in

, indicating different modes; />

, representing the center frequency of different modes; />

is the Dirac function, K is the number of modes to be decomposed, e is the natural base, t is time, j is a complex number, />

Indicates the original signal frequency;

求解方程（1）和（2），引入Lagrange乘法算子

，将约束变分问题转变为非约束变分问题，得到增广Lagrange求解公式：Solving equations (1) and (2), introducing the Lagrange multiplication operator

, transform the constrained variational problem into an unconstrained variational problem, and obtain the augmented Lagrange solution formula:

式中：

为二次惩罚因子，作用是降低高斯噪声的千扰；In the formula:

is the secondary penalty factor, the function is to reduce the interference of Gaussian noise;

利用交替方向乘子（ADMM)迭代算法结合Parseval/Plancherel、傅里叶等距变换，优化得到各模态分量和中心频率，并搜寻增广Lagrange西数的鞍点，交替寻优迭代后的

，/>

和/>

的表达式如下，Using the Alternate Direction Multiplier (ADMM) iterative algorithm combined with Parseval/Plancherel and Fourier equidistant transform to optimize the modal components and center frequency, and search for the saddle point of the augmented Lagrange western number, and alternately optimize the iterative

, />

and />

The expression of is as follows,

，

,

式中：n 表示迭代次数，k 表示模态数量，

为分解得到的基函数，/>

为第k个基函数的中心频率，/>

表示原信号频率，λ为 lagrange 乘法算子，/>

表示噪声容限，arg min 表示使后面这个式子达到最小值时的变量的取值；In the formula: n represents the number of iterations, k represents the number of modes,

is the basis function obtained by decomposition, />

is the center frequency of the kth basis function, />

Represents the original signal frequency, λ is the lagrange multiplication operator, />

Indicates the noise tolerance, and arg min indicates the value of the variable when the following formula reaches the minimum value;

求解得：

，根据BCG频谱分析结果选择中心频率个数为7，分量0为呼吸分量，选择分量1、2组成BCG分量。Solved:

According to the BCG spectrum analysis results, the number of center frequencies is selected as 7,component 0 is the respiratory component, and components 1 and 2 are selected to form the BCG component.

进一步地，在搭建非接触心电合成模型步骤中，对健康心脏的心电信号ECG进行心拍分割具体步骤为：在长时ECG信号中检索R波后根据生物机理规律，R波前0.36秒后0.5秒为一次心跳的ECG图，以每心拍ECG起止点作为时间标定分割同步BCG信号，根据BCG各波的产生机制和心动周期事件的时长确定BCG各波位置，取ECG的R波后0.35-0.65s内最大峰值为BCG的J波，J波前最近的极小值、极大值分别为I、H波；J波后最近极小值、极大值为K、L波。Further, in the step of building the non-contact electrocardiogram synthesis model, the specific steps of heart beat segmentation for the ECG signal of the healthy heart are as follows: after the R wave is retrieved from the long-term ECG signal, according to the biological mechanism, the R wave is 0.36 seconds after the front 0.5 seconds is the ECG diagram of a heartbeat, and the start and end points of each heartbeat ECG are used as the time calibration to divide the synchronous BCG signal. The position of each BCG wave is determined according to the generation mechanism of each BCG wave and the duration of the cardiac cycle event. The maximum peak within 0.65s is the J wave of BCG, and the nearest minimum and maximum values before the J wave are I and H waves respectively; after the J wave, the nearest minimum and maximum values are K and L waves.

进一步地，在搭建非接触心电合成模型步骤中，对分割后信号进行质量评估具体为：心电信号的质量评估基于随机森林融合SQI频率偏度、R波峰值、SQI信号频谱占比、低频段噪声频谱占比、R波变异性、R波是否易检测六个指标，数据集选择Computing inCardiology Challenge 2011数据集，其中R波是否易检测是用滑动窗口寻峰值和双斜线找不可导点两种R波检测的结果差异评估，其结果优则标记为1，劣为0。Further, in the step of building the non-contact ECG synthesis model, the quality assessment of the segmented signal is specifically as follows: the quality assessment of the ECG signal is based on random forest fusion of SQI frequency skewness, R wave peak value, SQI signal spectrum ratio, low Frequency band noise spectrum ratio, R-wave variability, and whether R-wave is easy to detect are six indicators. The data set is Computing inCardiology Challenge 2011 data set, and whether R-wave is easy to detect is to use the sliding window to find the peak value and the double slash to find the non-conductive point Evaluation of the difference between the results of the two R-wave detections, if the result is good, it is marked as 1, and if the result is bad, it is marked as 0.

进一步地，在对心脏进行评估步骤中，基于非接触的心电图数据和实际测量的心冲击信号BCG，计算以下参数：R-J同射血前期，为R波与J波时间间隔，能够反映出心脏收缩力的变化快慢；RI间隔，R波与I波时间间隔，是心脏收缩能力的经典BCG测量值；IJ波振幅差反映了与主动脉血压差相关的心输出量CO的波动，j波对应心室瓣膜运动。Further, in the step of assessing the heart, the following parameters are calculated based on the non-contact electrocardiogram data and the actual measured cardiac shock signal BCG: R-J is the same as the pre-ejection period, which is the time interval between the R wave and the J wave, which can reflect the heart contraction The speed of force change; RI interval, the time interval between R wave and I wave, is the classic BCG measurement of cardiac contractility; IJ wave amplitude difference reflects the fluctuation of cardiac output CO related to aortic blood pressure difference, and j wave corresponds to ventricular valve movement.

进一步地，所述心脏实时无感监测综合评价方法还包括心血管综合症预测步骤，所述心血管综合症预测步骤具体为：基于R-J同射血前期建立对数回归模型，Further, the comprehensive evaluation method for real-time non-inductive cardiac monitoring also includes a cardiovascular syndrome prediction step, which specifically includes: establishing a logarithmic regression model based on R-J and pre-ejection period,

（7），

(7),

式中，PEP为R-J同射血前期，

为预测值，/>

为系数。In the formula, PEP is the pre-ejection period of RJ,

is the predicted value, />

is the coefficient.

进一步地，在搭建非接触心电合成模型步骤中，Further, in the step of building a non-contact electrocardiogram synthesis model,

在搭建非接触心电合成模型步骤中，采用质量好的数据段，搭建非接触心电合成模型，所述质量好的数据段包括分割后同一次心跳的电活动和BCG中的心脏收缩、舒张活动产生的运动分量motion，具体为：In the step of building the non-contact electrocardiosynthesis model, the non-contact electrocardiography model is built by using data segments with good quality, which include the electrical activity of the same heartbeat after segmentation and the systole and diastole in the BCG The motion component motion generated by the activity, specifically:

编码器将motion抽样编码为隐变量z，The encoder encodes motion samples into hidden variables z,

，解码器将隐变量z重构为重构数据，得到心脏电活动数据/>

，

, the decoder reconstructs the latent variable z into reconstructed data, and obtains the cardiac electrical activity data/>

,

，式中，motion为心脏收缩、舒张运动信号，z为隐变量,

和ECG分别合成ECG和真实ECG，/>

为编码函数，/>

为解码函数，/>

、/>

为输入层到隐藏层权重，/>

、/>

为隐藏层到输出层权重，/>

为sigmoid函数。

, where motion is the systolic and diastolic motion signal, z is the hidden variable,

Synthetic ECG and real ECG respectively, />

is the encoding function, />

is the decoding function, />

, />

For input layer to hidden layer weights, />

, />

is the weight from the hidden layer to the output layer, />

is the sigmoid function.

本发明的目的之二采用如下技术方案实现：Two of the purpose of the present invention adopts following technical scheme to realize:

一种心脏实时无感监测综合评价装置，包括A real-time sensorless monitoring comprehensive evaluation device for the heart, comprising

柔性心脏带，用于获取待监测人体的原始体动数据；Flexible heart belt, used to obtain the raw body motion data of the human body to be monitored;

分析模块，用于根据上述心脏实时无感监测综合评价方法监测并综合评价心脏。The analysis module is used for monitoring and comprehensively evaluating the heart according to the above-mentioned comprehensive evaluation method for real-time non-inductive monitoring of the heart.

本发明的目的之三采用如下技术方案实现：Three of the purpose of the present invention adopts following technical scheme to realize:

一种心脏实时无感监测综合评价电子设备，包括An electronic device for comprehensive evaluation of heart real-time sensorless monitoring, including

处理器；processor;

存储器，所述存储器与所述处理器通信连接；a memory communicatively connected to the processor;

所述存储器存储有可被所述处理器执行的指令，所述指令被所述处理器执行以实现上述心脏实时无感监测综合评价方法。The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to realize the above-mentioned comprehensive evaluation method for real-time cardiac non-sensing monitoring.

相比现有技术，本发明心脏实时无感监测综合评价方法通过体征采集、对原始体动数据进行处理、搭建非接触心电合成模型、合成非接触的心电图数据以及对心脏进行评估等步骤，根据心脏机械运动信号和心脏电信号属同源不同表征，合成了实时、高精度的非接触心电图；基于所得非接触心电图和采集心冲击图，对心脏进行了多信号监测的基础上，进行综合评估和心血管疾病综合症的预测，基于该方法的合成的心电图能够有效地保留心脏电生理功能信息、生物差异性以及病理信息；突破了现有心血管移动评估设备采集与评估冲突的困境，实现了在单个设备上从泵生理和电生理两个维度无感评估心脏生理功能；基于合成ECG和BCG的电机械隔离时间PEP的对数模型实现了动态血压追踪、电生理、泵生理参数的准确计算；该模型基于非接触采集的BCG信号，其检测方式更方便、安全；同时相较于采集其他模态数据的非接触设备，本公开实施例受到的干扰小，易充电。Compared with the prior art, the comprehensive evaluation method for real-time non-inductive heart monitoring of the present invention collects signs, processes raw body motion data, builds a non-contact ECG synthesis model, synthesizes non-contact ECG data, and evaluates the heart. According to the homology and different characteristics of cardiac mechanical motion signal and cardiac electrical signal, a real-time, high-precision non-contact electrocardiogram was synthesized; based on the obtained non-contact electrocardiogram and collected shockcardiogram, multi-signal monitoring of the heart was carried out, and comprehensive Evaluation and prediction of cardiovascular disease syndrome, the synthetic electrocardiogram based on this method can effectively retain cardiac electrophysiological function information, biological differences and pathological information; break through the dilemma of existing cardiovascular mobile evaluation equipment acquisition and evaluation conflicts, realize Non-inductive assessment of cardiac physiological function from two dimensions of pump physiology and electrophysiology on a single device; the logarithmic model of electromechanical isolation time PEP based on synthetic ECG and BCG realizes accurate dynamic blood pressure tracking, electrophysiology, and pump physiological parameters calculation; the model is based on non-contact collected BCG signals, and its detection method is more convenient and safe; at the same time, compared with non-contact devices that collect other modal data, the embodiment of the present disclosure suffers less interference and is easy to charge.

附图说明Description of drawings

图1为本发明心脏实时无感监测综合评价方法的流程图；Fig. 1 is the flow chart of heart real-time sensorless monitoring comprehensive evaluation method of the present invention;

图2为BCG信号分解结果；Figure 2 is the decomposition result of BCG signal;

图3为BCG信号和ECG信号分割结果；Fig. 3 is the segmentation result of BCG signal and ECG signal;

图4为非接触心电合成模型的网络架构；Fig. 4 is the network architecture of the non-contact electrocardiogram synthesis model;

图5为非接触心电合成模型合成的非接触的心电图。Fig. 5 is a non-contact electrocardiogram synthesized by the non-contact electrocardiogram synthesis model.

具体实施方式Detailed ways

下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

需要说明的是，当组件被称为“固定于”另一个组件，它可以直接在另一个组件上或者也可以存在另一中间组件，通过中间组件固定。当一个组件被认为是“连接”另一个组件，它可以是直接连接到另一个组件或者可能同时存在另一中间组件。当一个组件被认为是“设置于”另一个组件，它可以是直接设置在另一个组件上或者可能同时存在另一中间组件。本文所使用的术语“垂直的”、“水平的”、“左”、“右”以及类似的表述只是为了说明的目的。It should be noted that when a component is said to be "fixed" to another component, it may be directly on the other component or there may be another intermediate component through which it is fixed. When a component is said to be "connected" to another component, it may be directly connected to the other component or there may be another intermediate component at the same time. When a component is said to be "set on" another component, it may be set directly on the other component or there may be another intermediate component at the same time. The terms "vertical," "horizontal," "left," "right," and similar expressions are used herein for purposes of illustration only.

除非另有定义，本文所使用的所有的技术和科学术语与属于本发明的技术领域的技术人员通常理解的含义相同。本文中在本发明的说明书中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本发明。本文所使用的术语“及／或”包括一个或多个相关的所列项目的任意的和所有的组合。Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field of the invention. The terms used herein in the description of the present invention are for the purpose of describing specific embodiments only, and are not intended to limit the present invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

如图1所示，本发明心脏实时无感监测综合评价方法，包括以下步骤：As shown in Figure 1, the heart real-time sensorless monitoring comprehensive evaluation method of the present invention comprises the following steps:

体征采集：将柔性心脏带铺设在待监测人体胸部下方，利用柔性心脏带由于受到压力后产生微形变的特性，采集待监测人体的原始体动数据，并将原始体动数据无线传输至处理器；Sign collection: Lay the flexible heart belt under the chest of the human body to be monitored, and use the characteristics of the flexible heart belt to produce micro-deformation under pressure to collect the original body motion data of the human body to be monitored, and wirelessly transmit the original body motion data to the processor ;

对原始体动数据进行处理：处理器对原始体动数据中的心冲击信号BCG进行频谱分析，采用变分模态分解技术，提取BCG分量；Process the original body motion data: the processor performs spectrum analysis on the cardiac shock signal BCG in the original body motion data, and uses variational mode decomposition technology to extract the BCG component;

搭建非接触心电合成模型：对健康心脏的心电信号ECG进行心拍分割，并对分割后信号进行质量评估，采用质量好的分割后心电信号搭建非接触心电合成模型；Build a non-contact ECG synthesis model: Segment the heart rate of the ECG signal ECG of a healthy heart, and evaluate the quality of the segmented signal, and build a non-contact ECG synthesis model by using the segmented ECG signal with good quality;

合成非接触的心电图数据：将BCG分量输入至非接触心电合成模型，使用深度学习算法学习电-机械运动之间的非线性映射关系，合成非接触的心电图数据；Synthesize non-contact ECG data: input BCG components into the non-contact ECG synthesis model, use deep learning algorithm to learn the nonlinear mapping relationship between electrical-mechanical motion, and synthesize non-contact ECG data;

对心脏进行评估：基于非接触的心电图数据和实际测量的心冲击信号BCG，对人体从电生理、泵生理角度进行评估。Evaluate the heart: Based on the non-contact ECG data and the actual measured cardiac shock signal BCG, the human body is evaluated from the perspective of electrophysiology and pump physiology.

具体的，在体征采集步骤中，柔性心脏带包括压电传感器、前置电荷放大电路二阶带通滤波电路以及增益放大电路。压电传感器与前置电荷放大电路二阶带通滤波电路以及增益放大电路电连接，压电传感器由聚偏氟乙烯制成，聚偏氟乙烯(PolyvinylideneFluoride，PVDF)材料经拉伸后表面出现极化，受到压力后产生微形变，并在极化面产生极性与运动方向一致的瞬变电荷。原始体动数据经WIFI通讯模块将信号传输到计算机进行后续处理与可视化。采集待监测人体卧床时3种状态的原始体动数据，3种状态分别为：静息状态，待监测人体被要求统一以双手自然垂在身侧、保持均匀呼吸的姿势；运动后状态，待监测人体被要求进行若干次原地下蹲运动后卧床；睡眠状态，待监测人体处于睡眠中。在本实施例中，原地下蹲次数为15次。实验结束后人为筛选较稳定段。单次实验长度4分钟，受试者被要求在实验开始和结束时咳嗽三次，在两个设备的波形上引入明显的运动伪影，作为开始和结束的标志。Specifically, in the sign collection step, the flexible heart belt includes a piezoelectric sensor, a pre-charge amplifier circuit, a second-order band-pass filter circuit, and a gain amplifier circuit. The piezoelectric sensor is electrically connected to the second-order band-pass filter circuit and the gain amplifier circuit of the pre-charge amplifier circuit. The piezoelectric sensor is made of polyvinylidene fluoride. After being subjected to pressure, micro-deformation occurs, and a transient charge with a polarity consistent with the direction of motion is generated on the polarized surface. The original body motion data is transmitted to the computer through the WIFI communication module for subsequent processing and visualization. Collect the original body movement data of three states when the human body to be monitored is lying in bed. The three states are: the resting state, the human body to be monitored is required to hang its hands naturally by the side and maintain a uniform breathing posture; the state after exercise, the waiting state The monitoring human body is required to lie in bed after several times of original squat exercise; in the sleeping state, the human body to be monitored is in sleep. In this embodiment, the original number of times of squatting down is 15 times. After the experiment, the more stable segment was artificially screened. The length of a single experiment was 4 minutes, and the subjects were asked to cough three times at the beginning and end of the experiment, introducing obvious motion artifacts on the waveforms of both devices as a sign of the beginning and end.

具体的，在对原始体动数据进行处理步骤中，柔性心脏带采集的原始数据包括心冲击信号（BCG）、呼吸引起的体动和环境噪声。心冲击信号产生过程分析可知，其又包含心肌收缩舒张、运动和心脏内血流、瓣膜等机械运动。通过对BCG进行频谱分析，BCG信号分布范围为1-10Hz，呼吸信号0.21-0.9Hz。针对该信号特征特点，使用改进的变分模态分解（Variational Mode Decomposition，VMD）技术。根据BCG频谱分析结果选择中心频率个数为7，提取到呼吸和BCG分量。改进的变分模态分解公式如下：Specifically, in the step of processing the raw body motion data, the raw data collected by the flexible heart belt includes a heart attack signal (BCG), body motion caused by respiration, and environmental noise. The analysis of the generation process of the cardiac shock signal shows that it also includes myocardial contraction and relaxation, exercise, and mechanical movements such as blood flow in the heart and valves. By analyzing the spectrum of BCG, the distribution range of BCG signal is 1-10Hz, and the respiratory signal is 0.21-0.9Hz. According to the characteristics of the signal, the improved variational mode decomposition (Variational Mode Decomposition, VMD) technology is used. According to the results of BCG spectrum analysis, the number of center frequencies is selected as 7, and the breath and BCG components are extracted. The improved variational mode decomposition formula is as follows:

（1）

(1)

（2）

(2)

其中

,表示不同模态；/>

，表示不同模态中心频率；/>

为狄拉克函数，K为需要分解的模态个数，e为自然底数，t为时间，j表示复数，/>

表示原信号频率；in

, indicating different modes; />

, representing the center frequency of different modes; />

is the Dirac function, K is the number of modes to be decomposed, e is the natural base, t is time, j is a complex number, />

Indicates the original signal frequency;

求解方程（1）和（2），引入Lagrange乘法算子

，将约束变分问题转变为非约束变分问题，得到增广lagrange求解公式：Solving equations (1) and (2), introducing the Lagrange multiplication operator

, transform the constrained variational problem into an unconstrained variational problem, and obtain the augmented lagrange solution formula:

，式中：/>

为二次惩罚因子，作用是降低高斯噪声的千扰；

, where: />

is the secondary penalty factor, the function is to reduce the interference of Gaussian noise;

解该约束问题可通过，交替方向乘子法ADMM来求解，其思想简述为固定另外两个变量，更新其中一个变量，如下：Solving this constraint problem can be solved by the Alternating Direction Multiplier Method ADMM. The idea is simply to fix the other two variables and update one of them, as follows:

利用交替方向乘子（ADMM)迭代算法结合Parseval/Plancherel、傅里叶等距变换，优化得到各模态分量和中心频率，并搜寻增广Lagrange西数的鞍点，交替寻优迭代后的

，/>

和/>

的表达式如下，Using the Alternate Direction Multiplier (ADMM) iterative algorithm combined with Parseval/Plancherel and Fourier equidistant transform to optimize the modal components and center frequency, and search for the saddle point of the augmented Lagrange western number, and alternately optimize the iterative

, />

and />

The expression of is as follows,

，

,

式中：n 表示迭代次数，k 表示模态数量，

为分解得到的基函数，/>

为第k个基函数的中心频率，/>

表示原信号频率，λ为 lagrange 乘法算子，/>

表示噪声容限，arg min 表示使后面这个式子达到最小值时的变量的取值；In the formula: n represents the number of iterations, k represents the number of modes,

is the basis function obtained by decomposition, />

is the center frequency of the kth basis function, />

Represents the original signal frequency, λ is the lagrange multiplication operator, />

Indicates the noise tolerance, and arg min indicates the value of the variable when the following formula reaches the minimum value;

求解得：Solved:

，

,

根据BCG频谱分析结果选择中心频率个数为7，如附图2所示。分量0为呼吸分量，选择分量1、2组成BCG分量。According to the BCG spectrum analysis results, the number of center frequencies is selected as 7, as shown in Figure 2.Component 0 is the respiratory component, and components 1 and 2 are selected to form the BCG component.

具体的，在搭建非接触心电合成模型步骤中，对健康心脏的心电信号ECG进行心拍分割具体步骤为：在长时ECG信号中检索R波后根据生物机理规律，R波前0.36秒后0.5秒为一次心跳的ECG图，以每心拍ECG起止点作为时间标定分割同步BCG信号，根据BCG各波的产生机制和心动周期事件的时长确定BCG各波位置，取ECG的R波后0.35-0.65s内最大峰值为BCG的J波，J波前最近的极小值、极大值分别为I、H波；J波后最近极小值、极大值为K、L波。ECG的最终预处理结果如图3所示。从心脏电收缩开始识别的第一、第二和第三极大值处被命名为H、J、L波，极小值为I、K波。Specifically, in the step of building a non-contact electrocardiogram synthesis model, the specific steps for heart beat segmentation of the ECG signal ECG of a healthy heart are as follows: after the R wave is retrieved from the long-term ECG signal, according to the law of biological mechanism, after the R wave is 0.36 seconds before 0.5 seconds is the ECG diagram of a heartbeat, and the start and end points of each heartbeat ECG are used as the time calibration to divide the synchronous BCG signal. The position of each BCG wave is determined according to the generation mechanism of each BCG wave and the duration of the cardiac cycle event. The maximum peak within 0.65s is the J wave of BCG, and the nearest minimum and maximum values before the J wave are I and H waves respectively; after the J wave, the nearest minimum and maximum values are K and L waves. The final preprocessing results of ECG are shown in Fig. 3. The first, second, and third maxima identified from the onset of electrical contraction are named H, J, and L waves, and the minima are I, K waves.

对分割后信号进行质量评估具体为：心电信号的质量评估基于随机森林融合SQI频率偏度、R波峰值、SQI信号频谱占比、低频段噪声频谱占比、R波变异性、R波是否易检测六个指标，数据集选择Computing in Cardiology Challenge 2011数据集，其中R波是否易检测是用滑动窗口寻峰值和双斜线找不可导点两种R波检测的结果差异评估，其结果优则标记为1，劣为0。The quality assessment of the segmented signal is specifically: the quality assessment of the ECG signal is based on random forest fusion of SQI frequency skewness, R wave peak value, SQI signal spectrum proportion, low frequency noise spectrum proportion, R wave variability, whether R wave The six indicators are easy to detect, and the data set is Computing in Cardiology Challenge 2011 data set. The R wave is easy to detect. It is marked as 1 and bad as 0.

采用质量好的分割后心电信号及对应的同步BCG分量搭建非接触心电合成模型如图4所示，具体为：The non-contact ECG synthesis model is built by using good quality segmented ECG signals and corresponding synchronous BCG components, as shown in Figure 4, specifically:

在搭建非接触心电合成模型步骤中，采用质量好的数据段（包括分割后同一次心跳的电活动和BCG中的心脏心脏收缩、舒张活动产生的运动分量motion）搭建非接触心电合成模型。具体为：In the step of building a non-contact ECG model, a non-contact ECG model is built using good-quality data segments (including the electrical activity of the same heartbeat after segmentation and the motion component motion generated by the cardiac systole and diastolic activity in BCG) . Specifically:

编码器将BCG中心脏收缩、舒张运动信号motion抽样编码为隐变量z，The encoder encodes the motion sampling of the systolic and diastolic motion signals in the BCG into a hidden variable z,

解码器将隐变量z重构为重构数据，得到心脏电活动数据

。The decoder reconstructs the hidden variable z into reconstructed data to obtain the heart electrical activity data

.

，式中，motion为心脏收缩、舒张运动信号，z为隐变量,/>

和ECG分别合成ECG和真实ECG，/>

为编码函数，/>

为解码函数，/>

、/>

为输入层到隐藏层权重，/>

、/>

为隐藏层到输出层权重，/>

为sigmoid函数。

, where motion is the systolic and diastolic motion signal, z is the hidden variable, />

Synthetic ECG and real ECG respectively, />

is the encoding function, />

is the decoding function, />

, />

For input layer to hidden layer weights, />

, />

is the weight from the hidden layer to the output layer, />

is the sigmoid function.

优化目标：

，其中dist为L2损失

。optimize the target:

, where dist is the L2 loss

.

在对心脏进行评估步骤中，基于非接触的心电图数据和实际测量的心冲击信号BCG，计算以下参数：In the step of evaluating the heart, the following parameters are calculated based on the non-contact ECG data and the actually measured shock signal BCG:

特征名称feature name描述describeR_J(PEP)R_J(PEP)R波与J波时间间隔R wave and J wave time intervalRJ AmpRJ AmpR波与J波振幅差R wave and J wave amplitude differenceR_IR_IR波与I波时间间隔R wave and I wave time intervalR_I AmpR_I AmpR波与I波振幅差R wave and I wave amplitude differenceI_JI_JI波与J波时间间隔Time interval between I wave and J waveI_J AmpI_J AmpI波与J波振幅差I wave and J wave amplitude difference

R-J同射血前期，为R波与J波时间间隔，能够反映出心脏收缩力的变化快慢；RI间隔，R波与I波时间间隔，是心脏收缩能力的经典BCG测量值；IJ波振幅差反映了与主动脉血压差相关的心输出量CO的波动，j波对应心室瓣膜运动。R-J is the same as the pre-ejection period, which is the time interval between R wave and J wave, which can reflect the change speed of cardiac contractility; RI interval, the time interval between R wave and I wave, is the classic BCG measurement value of cardiac contractility; IJ wave amplitude difference Reflecting fluctuations in cardiac output CO related to the aortic pressure difference, the j-wave corresponds to ventricular valve motion.

心脏实时无感监测综合评价方法还包括心血管综合症预测步骤，心冲击信号BCG描述血液动力学变化，重构ECG信号跟踪心脏激动传导的变化。可以证明使用这种非接触设备可以实现家庭场景中持续跟踪人体人体心脏动力学的变化，并分析房颤对血流动力学的影响造成的远期血流栓塞风险。The comprehensive evaluation method of cardiac real-time non-inductive monitoring also includes the step of predicting cardiovascular syndrome, cardiac shock signal BCG to describe hemodynamic changes, reconstructing ECG signal to track changes in cardiac excitation conduction. It can be proved that the use of this non-contact device can continuously track changes in human heart dynamics in the home scene, and analyze the long-term blood embolism risk caused by the impact of atrial fibrillation on hemodynamics.

心血管综合症预测步骤具体为：基于R-J同射血前期建立对数回归模型，The specific steps for predicting cardiovascular syndrome are as follows: establishing a logarithmic regression model based on R-J and pre-ejection period,

（7）

(7)

式中，PEP为R-J同射血前期，

为预测值，/>

为系数。In the formula, PEP is the pre-ejection period of RJ,

is the predicted value, />

is the coefficient.

本申请基于所得非接触心电图和采集心冲击图，对心脏进行了多信号监测的基础上，进行综合评估和心血管疾病综合症的预测，基于该方法的合成的心电图能够有效地保留心脏电生理功能信息、生物差异性以及病理信息；突破了现有心血管移动评估设备采集与评估冲突的困境，实现了在单个设备上从泵生理和电生理两个维度无感评估心脏生理功能；基于合成ECG和BCG的电机械隔离时间PEP的对数模型实现了动态血压追踪、电生理、泵生理参数的准确计算；该模型基于非接触采集的BCG信号，其检测方式更方便、安全；同时相较于采集其他模态数据的非接触设备，本公开实施例受到的干扰小，易充电。This application is based on the obtained non-contact electrocardiogram and collected ballistocardiogram, and on the basis of multi-signal monitoring of the heart, comprehensive assessment and prediction of cardiovascular disease syndrome are carried out. The synthetic electrocardiogram based on this method can effectively preserve the cardiac electrophysiological Functional information, biological differences, and pathological information; break through the dilemma of acquisition and evaluation conflicts of existing cardiovascular mobile evaluation equipment, and realize non-inductive evaluation of cardiac physiological function from two dimensions of pump physiology and electrophysiology on a single device; based on synthetic ECG The logarithmic model of the electromechanical isolation time PEP of BCG realizes the accurate calculation of ambulatory blood pressure tracking, electrophysiology, and pump physiological parameters; the model is based on non-contact collected BCG signals, and its detection method is more convenient and safe; at the same time, compared with For non-contact devices that collect other modal data, the embodiment of the present disclosure suffers little interference and is easy to charge.

以上实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进演变，都是依据本发明实质技术对以上实施例做的等同修饰与演变，这些都属于本发明的保护范围。The above examples only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as limiting the scope of the patent for the invention. It should be pointed out that for those skilled in the art, without departing from the concept of the present invention, several modifications and improvements can be made, all of which are equivalent modifications made to the above embodiments based on the substantive technology of the present invention and evolution, these all belong to the protection scope of the present invention.

Claims (11)

Translated fromChinese

1.一种心脏实时无感监测综合评价方法，其特征在于，包括以下步骤：1. A real-time non-inductive monitoring comprehensive evaluation method for the heart, characterized in that, comprising the following steps:体征采集：将柔性心脏带铺设在待监测人体胸部下方，利用柔性心脏带由于受到压力后产生微形变的特性，采集待监测人体的原始体动数据，并将原始体动数据无线传输至处理器；Sign collection: Lay the flexible heart belt under the chest of the human body to be monitored, and use the characteristics of the flexible heart belt to produce micro-deformation under pressure to collect the original body motion data of the human body to be monitored, and wirelessly transmit the original body motion data to the processor ;对原始体动数据进行处理：处理器对原始体动数据中的心冲击信号BCG进行频谱分析，采用变分模态分解技术，提取BCG分量；Process the original body motion data: the processor performs spectrum analysis on the cardiac shock signal BCG in the original body motion data, and uses variational mode decomposition technology to extract the BCG component;搭建非接触心电合成模型：对健康心脏的心电信号ECG进行心拍分割，并对分割后信号进行质量评估，采用质量好的分割后心电信号搭建非接触心电合成模型；Build a non-contact ECG synthesis model: Segment the heart rate of the ECG signal ECG of a healthy heart, and evaluate the quality of the segmented signal, and build a non-contact ECG synthesis model by using the segmented ECG signal with good quality;合成非接触的心电图数据：将BCG分量输入至非接触心电合成模型，使用深度学习算法学习电-机械运动之间的非线性映射关系，合成非接触的心电图数据；Synthesize non-contact ECG data: input BCG components into the non-contact ECG synthesis model, use deep learning algorithm to learn the nonlinear mapping relationship between electrical-mechanical motion, and synthesize non-contact ECG data;对心脏进行评估：基于非接触的心电图数据和实际测量的心冲击信号BCG，对人体从电生理、泵生理角度进行评估。Evaluate the heart: Based on the non-contact ECG data and the actual measured cardiac shock signal BCG, the human body is evaluated from the perspective of electrophysiology and pump physiology.2.根据权利要求1所述的心脏实时无感监测综合评价方法，其特征在于：在所述体征采集步骤中，所述柔性心脏带包括压电传感器，所述压电传感器由聚偏氟乙烯制成。2. The heart real-time non-inductive monitoring comprehensive evaluation method according to claim 1, characterized in that: in the sign collection step, the flexible heart belt includes a piezoelectric sensor, and the piezoelectric sensor is made of polyvinylidene fluoride production.3.根据权利要求1所述的心脏实时无感监测综合评价方法，其特征在于：在所述体征采集步骤中，采集待监测人体卧床时3种状态的原始体动数据，3种状态分别为：静息状态，待监测人体被要求统一以双手自然垂在身侧、保持均匀呼吸的姿势；运动后状态，待监测人体被要求进行若干次原地下蹲运动后卧床；睡眠状态，待监测人体处于睡眠中。3. The heart real-time sensorless monitoring comprehensive evaluation method according to claim 1, characterized in that: in the sign collection step, collect the original body motion data of 3 states when the human body to be monitored is lying in bed, and the 3 states are respectively : In the resting state, the human body to be monitored is required to hang its hands naturally at the side and maintain a posture of breathing evenly; in the post-exercise state, the human body to be monitored is required to lie in bed after performing several squat exercises in the same position; in the sleeping state, the human body to be monitored is asleep.4.根据权利要求1所述的心脏实时无感监测综合评价方法，其特征在于：在对原始体动数据进行处理步骤中，采用变分模态分解技术，提取BCG分量具体为：4. The heart real-time sensorless monitoring comprehensive evaluation method according to claim 1, characterized in that: in the processing step of the original body motion data, the variational mode decomposition technique is adopted to extract the BCG component as follows:变分模态分解公式为：The variational mode decomposition formula is:

（1）

(1)

（2）

(2)其中

,表示不同模态；/>

，表示不同模态中心频率；/>

为狄拉克函数，K为需要分解的模态个数，e为自然底数，t为时间，j 表示复数，/>

表示原信号频率；求解方程（1）和（2），引入Lagrange乘法算子/>

，将约束变分问题转变为非约束变分问题，得到增广Lagrange求解公式：in

, indicating different modes; />

, representing the center frequency of different modes; />

is the Dirac function, K is the number of modes to be decomposed, e is the natural base, t is time, j is a complex number, />

Represents the original signal frequency; solve equations (1) and (2), and introduce Lagrange multiplication operator/>

, transform the constrained variational problem into an unconstrained variational problem, and obtain the augmented Lagrange solution formula:

（3），

(3),式中：

为二次惩罚因子，作用是降低高斯噪声的干扰；In the formula:

It is a secondary penalty factor, which is used to reduce the interference of Gaussian noise;利用交替方向乘子（ADMM)迭代算法结合Parseval/Planchered、傅里叶等距变换，优化得到各模态分量和中心频率，并搜寻增广Lagrange函数的鞍点，交替寻优迭代后的

，/>

和/>

的表达式如下，Using the Alternate Direction Multiplier (ADMM) iterative algorithm combined with Parseval/Planchered and Fourier equidistant transform, optimize the modal components and center frequencies, and search for the saddle point of the augmented Lagrange function, and alternately optimize the iterative

, />

and />

The expression of is as follows,

（4）

(4)

（5）

(5)

（6）

(6)式中：n 表示迭代次数，k 表示模态数量，

为分解得到的基函数，/>

为第k个每个基函数的中心频率，f表示原信号频率，λ为 lagrange 乘法算子，/>

表示噪声容限，argmin 表示使后面这个式子达到最小值时的变量的取值；In the formula: n represents the number of iterations, k represents the number of modes,

is the basis function obtained by decomposition, />

is the center frequency of each k-th basis function,f represents the original signal frequency, λ is the lagrange multiplication operator, />

Indicates the noise tolerance, and argmin indicates the value of the variable when the following formula reaches the minimum value;求解得：

，Solved:

,

，

,根据BCG频谱分析结果选择中心频率个数为7，分量0为呼吸分量，选择分量1、2组成BCG分量。According to the results of BCG spectrum analysis, the number of center frequencies is selected as 7, component 0 is the respiratory component, and components 1 and 2 are selected to form the BCG component.5.根据权利要求1所述的心脏实时无感监测综合评价方法，其特征在于：在搭建非接触心电合成模型步骤中，对健康心脏的心电信号ECG进行心拍分割具体步骤为：在长时ECG信号中检索R波后根据生物机理规律，R波前0.36秒后0.5秒为一次心跳的ECG图，以每心拍ECG起止点作为时间标定分割同步BCG信号，根据BCG各波的产生机制和心动周期事件的时长确定BCG各波位置，取ECG的R波后0.35-0.65s内最大峰值为BCG的J波，J波前最近的极小值、极大值分别为I、H波；J波后最近极小值、极大值为K、L波。5. The heart real-time non-inductive monitoring comprehensive evaluation method according to claim 1, characterized in that: in the step of building a non-contact electrocardiogram synthesis model, the specific steps of heart beat segmentation to the electrocardiographic signal ECG of a healthy heart are: After searching the R wave in the hourly ECG signal, according to the law of biological mechanism, 0.36 seconds before the R wave and 0.5 seconds after the R wave is the ECG map of a heartbeat, and the ECG start and end points of each heart beat are used as the time calibration to divide the synchronous BCG signal. According to the generation mechanism of BCG waves and The duration of cardiac cycle events determines the position of each wave of BCG. The maximum peak value within 0.35-0.65s after the R wave of ECG is taken as the J wave of BCG, and the nearest minimum and maximum values before the J wave are I and H waves respectively; J The nearest minimum value and maximum value after the wave are K and L waves.6.根据权利要求5所述的心脏实时无感监测综合评价方法，其特征在于：在搭建非接触心电合成模型步骤中，对分割后信号进行质量评估具体为：心电信号的质量评估基于随机森林融合SQI频率偏度、R波峰值、SQI信号频谱占比、低频段噪声频谱占比、R波变异性、R波是否易检测六个指标，数据集选择Computing in Cardiology Challenge 2011数据集，其中R波是否易检测是用滑动窗口寻峰值和双斜线找不可导点两种R波检测的结果差异评估，其结果优则标记为1，劣为0。6. The heart real-time non-inductive monitoring comprehensive evaluation method according to claim 5, characterized in that: in the step of setting up a non-contact electrocardiogram synthesis model, the quality evaluation of the signal after segmentation is specifically: the quality evaluation of the electrocardiographic signal is based on The random forest combines SQI frequency skewness, R-wave peak value, SQI signal spectrum proportion, low-frequency noise spectrum proportion, R-wave variability, and whether R-wave is easy to detect. The data set is Computing in Cardiology Challenge 2011 data set. Whether the R wave is easy to detect is evaluated by the difference between the results of the two R wave detections using the sliding window to find the peak value and the double slash to find the non-conductive point. The result is marked as 1 if the result is good, and 0 if it is bad.7.根据权利要求5所述的心脏实时无感监测综合评价方法，其特征在于：在对心脏进行评估步骤中，基于非接触的心电图数据和实际测量的心冲击信号BCG，计算以下参数：R-J同射血前期，为R波与J波时间间隔，能够反映出心脏收缩力的变化快慢；RI间隔，R波与I波时间间隔，是心脏收缩能力的经典BCG测量值；IJ波振幅差反映了与主动脉血压差相关的心输出量CO的波动，j波对应心室瓣膜运动。7. The heart real-time sensorless monitoring comprehensive evaluation method according to claim 5, characterized in that: in the step of evaluating the heart, the following parameters are calculated based on the non-contact electrocardiogram data and the heart shock signal BCG of the actual measurement: R-J Same as the pre-ejection period, it is the time interval between R wave and J wave, which can reflect the change speed of cardiac contractility; RI interval, the time interval between R wave and I wave, is the classic BCG measurement value of cardiac contractility; IJ wave amplitude difference reflects Fluctuations in cardiac output CO associated with aortic blood pressure differences are accounted for, and the j-wave corresponds to ventricular valve motion.8.根据权利要求7所述的心脏实时无感监测综合评价方法，其特征在于：所述心脏实时无感监测综合评价方法还包括心血管综合症预测步骤，所述心血管综合症预测步骤具体为：基于R-J同射血前期建立对数回归模型，8. The heart real-time sensorless monitoring comprehensive evaluation method according to claim 7, characterized in that: the heart real-time sensorless monitoring comprehensive evaluation method also includes a cardiovascular syndrome prediction step, and the cardiovascular syndrome prediction step is specifically To: Establish a logarithmic regression model based on the R-J same pre-ejection period,

（7），

(7),式中，PEP为R-J同射血前期，

为预测值，/>

为系数。In the formula, PEP is the pre-ejection period of RJ,

is the predicted value, />

is the coefficient.9.根据权利要求1所述的心脏实时无感监测综合评价方法，其特征在于：在搭建非接触心电合成模型步骤中，采用质量好的数据段，搭建非接触心电合成模型，所述质量好的数据段包括分割后同一次心跳的电活动和BCG中的心脏收缩、舒张活动产生的运动分量motion，具体为：9. The heart real-time sensorless monitoring comprehensive evaluation method according to claim 1, characterized in that: in the step of setting up the non-contact electrocardiosynthesis model, the data segment with good quality is used to build the non-contact electrocardiography model, the The data segment with good quality includes the electrical activity of the same heartbeat after segmentation and the motion component motion generated by the systolic and diastolic activities in BCG, specifically:编码器将motion抽样编码为隐变量z，The encoder encodes motion samples into hidden variables z,

，解码器将隐变量z重构为重构数据，得到心脏电活动数据

，

, the decoder reconstructs the latent variable z into reconstructed data, and obtains the cardiac electrical activity data

,

，式中，motion为心脏收缩、舒张运动信号，z为隐变量,/>

和ECG分别合成ECG和真实ECG，/>

为编码函数，/>

为解码函数，/>

、/>

为输入层到隐藏层权重，/>

、/>

为隐藏层到输出层权重，/>

为sigmoid函数。

, where motion is the systolic and diastolic motion signal, z is the hidden variable, />

Synthetic ECG and real ECG respectively, />

is the encoding function, />

is the decoding function, />

, />

For input layer to hidden layer weights, />

, />

is the weight from the hidden layer to the output layer, />

is the sigmoid function.10.一种心脏实时无感监测综合评价装置，其特征在于：包括10. A real-time non-inductive monitoring comprehensive evaluation device for the heart, characterized in that: comprising柔性心脏带，用于获取待监测人体的原始体动数据；Flexible heart belt, used to obtain the raw body motion data of the human body to be monitored;分析模块，用于根据权利要求1-9任意一项所述的心脏实时无感监测综合评价方法监测并综合评价心脏。The analysis module is used for monitoring and comprehensively evaluating the heart according to the comprehensive evaluation method for real-time non-inductive monitoring of the heart according to any one of claims 1-9.11.一种心脏实时无感监测综合评价电子设备，其特征在于：包括11. An electronic device for comprehensive evaluation of heart real-time non-inductive monitoring, characterized in that: comprising处理器；processor;存储器，所述存储器与所述处理器通信连接；a memory communicatively connected to the processor;所述存储器存储有可被所述处理器执行的指令，所述指令被所述处理器执行以实现权利要求1-9任意一项所述的心脏实时无感监测综合评价方法。The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to implement the comprehensive evaluation method for real-time cardiac non-sensing monitoring according to any one of claims 1-9.

Images