Method, device and computer equipment for evaluating quality of ballistocardiogram signal

Abstract

The present invention relates to the technical field of signal processing, and in particular to a quality assessment method for a ballistocardiogram signal. The method analyzes the characteristics of the ballistocardiogram signal, extracts an initial feature vector of the ballistocardiogram signal, performs shallow and deep information mining on the initial feature vector, obtains direct features and potential features of the ballistocardiogram signal that can affect quality assessment, improves the accuracy of signal quality assessment results, and thus improves the accuracy of ballistocardiogram signal extraction.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a quality assessment method, device, computer equipment and storage medium for a ballistocardiogram signal. Background Art

[0002] Ballistocardiogram (BCG) describes the weak vibration of the human body caused by the heart's ejection process, and has been proven to be an effective way to monitor the heart. BCG signal is a weak force signal that can be converted into an electrical signal through a pressure sensor without direct contact, thus achieving non-intrusive collection of the user's vital signs.

[0003] At present, the ballistocardiogram signal is mainly obtained through non-contact vital signs monitoring. The acquisition equipment does not need to be in direct contact with the subject's body. Its comfort and ease of operation make it have great advantages in individualized long-term home health monitoring. However, due to the weakness of the ballistocardiogram signal, its morphology is also easily disturbed. In the actual application scenarios of the above acquisition method, the signal will be disturbed by such interferences as breathing, motion artifacts, distance from the sensor, environmental noise, etc., resulting in inconsistent monitoring signal quality, and the same analysis and application cannot be performed. In addition, due to the individual differences of the ballistocardiogram signal, various influencing factors inside and outside the user's body in different populations will cause the ballistocardiogram morphology to change. Therefore, quality control of the ballistocardiogram signal and selection of relatively better quality data from the signal are of great significance for subsequent data processing and analysis. Summary of the invention

[0004] Based on this, the purpose of the present invention is to provide a quality assessment method, device, computer equipment and storage medium for a ballistocardiogram signal, by analyzing the characteristics of the ballistocardiogram signal, extracting the initial feature vector of the ballistocardiogram signal, performing shallow and deep information mining on the initial feature vector, obtaining direct features and potential features of the ballistocardiogram signal that can affect the quality assessment, improving the accuracy of the signal quality assessment results, and thus improving the accuracy of the extraction of the ballistocardiogram signal.

[0005] In a first aspect, an embodiment of the present application provides a method for evaluating the quality of a ballistocardiogram signal, comprising the following steps:

[0006] Obtaining a ballistocardiogram signal to be detected and a preset quality assessment model, dividing the ballistocardiogram signal to be detected into a plurality of ballistocardiogram unit signals to be detected, and obtaining a plurality of ballistocardiogram unit signals to be detected corresponding to the ballistocardiogram signal to be detected, wherein the quality assessment model includes a feature extraction module, a feature processing module and a quality assessment module;

[0007] Inputting a plurality of the to-be-detected ballistocardiogram unit signals into the feature extraction module for feature extraction, and obtaining initial feature vectors of the plurality of to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector;

[0008] Inputting the initial feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals into the feature processing module for feature processing, and obtaining the hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals, wherein the hierarchical feature vectors include shallow feature vectors and deep feature vectors;

[0009] The hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals are input into the quality assessment module to obtain quality assessment results of the plurality of the to-be-detected ballistocardiogram unit signals, and based on the quality assessment results, a plurality of target ballistocardiogram unit signals are extracted from the plurality of the to-be-detected ballistocardiogram unit signals.

[0010] In a second aspect, an embodiment of the present application provides a quality assessment device for a ballistocardiogram signal, comprising:

[0011] A data acquisition module, used for obtaining a ballistocardiogram signal to be detected and a preset quality assessment model, dividing the ballistocardiogram signal to be detected into a plurality of ballistocardiogram unit signals to be detected, and obtaining a plurality of ballistocardiogram unit signals to be detected corresponding to the ballistocardiogram signal to be detected, wherein the quality assessment model includes a feature extraction module, a feature processing module and a quality assessment module;

[0012] An initial feature extraction module, used for inputting a plurality of the to-be-detected ballistocardiogram unit signals into the feature extraction module for feature extraction, and obtaining initial feature vectors of the plurality of to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector;

[0013] A hierarchical feature extraction module, used for inputting the initial feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals into the feature processing module for feature processing, so as to obtain the hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals;

[0014] A signal quality assessment module is used to input the hierarchical feature vectors of the several ballistocardiogram unit signals to be detected into the quality assessment module, obtain quality assessment results of the several ballistocardiogram unit signals to be detected, and extract several target ballistocardiogram unit signals from the several ballistocardiogram unit signals to be detected based on the quality assessment results.

[0015] In a third aspect, an embodiment of the present application provides a computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for quality assessment of a ballistocardiogram signal as described in the first aspect when executing the computer program.

[0016] In a fourth aspect, an embodiment of the present application provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of the method for quality assessment of a ballistocardiogram signal as described in the first aspect.

[0017] In an embodiment of the present application, a quality assessment method, apparatus, computer equipment and storage medium for a ballistocardiogram signal are provided. The characteristics of the ballistocardiogram signal are analyzed to extract an initial feature vector of the ballistocardiogram signal. The initial feature vector is subjected to shallow and deep information mining to obtain direct features and potential features of the ballistocardiogram signal that can affect the quality assessment, thereby improving the accuracy of the signal quality assessment result and thus improving the accuracy of the extraction of the ballistocardiogram signal.

[0018] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic flow chart of a method for evaluating the quality of a ballistocardiogram signal provided in one embodiment of the present application;

[0020] Figure 2 A schematic diagram of S2 in the process of the method for evaluating the quality of a ballistocardiogram signal provided in the first embodiment of the present application;

[0021] Figure 3 A schematic diagram of S2 in the process of the method for evaluating the quality of a ballistocardiogram signal provided in the second embodiment of the present application;

[0022] Figure 4 A schematic diagram of S2 in the process of the method for evaluating the quality of a ballistocardiogram signal provided in the third embodiment of the present application;

[0023] Figure 5 A schematic diagram of S3 in the process of a method for evaluating the quality of a ballistocardiogram signal provided by an embodiment of the present application;

[0024] Figure 6 A schematic diagram of S4 in the process of a method for evaluating the quality of a ballistocardiogram signal provided by an embodiment of the present application;

[0025] Figure 7 A schematic diagram of the flow chart of S5 in the method for evaluating the quality of a ballistocardiogram signal provided by another embodiment of the present application;

[0026] Figure 8 A schematic diagram of the structure of a device for evaluating the quality of a ballistocardiogram signal provided in one embodiment of the present application;

[0027] Fig. 9 A schematic diagram of the structure of a computer device provided for one embodiment of the present application. DETAILED DESCRIPTION

[0028] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0029] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0030] It should be understood that, although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" / "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determination".

[0031] See also Figure 1 , Figure 1 A flow chart of a method for evaluating the quality of a ballistocardiogram signal provided in one embodiment of the present application, the method comprising the following steps:

[0032] S1: obtaining a ballistocardiogram signal to be detected and a preset quality assessment model, dividing the ballistocardiogram signal to be detected into a plurality of ballistocardiogram unit signals to be detected, and obtaining a plurality of ballistocardiogram unit signals to be detected corresponding to the ballistocardiogram signal to be detected.

[0033] The executor of the quality assessment method of the ballistocardiogram signal is an extraction device of the quality assessment method of the ballistocardiogram signal (hereinafter referred to as the extraction device). In an optional embodiment, the extraction device can be a computer device, a server, or a server cluster composed of multiple computer devices.

[0034] In an optional embodiment, the extraction device can obtain the heartbeat signal to be detected by querying a preset database. In another optional embodiment, the extraction device can use a piezoelectric sensor to obtain the user's human body micro-vibration signal without affecting the user's normal sleep, and convert the human body micro-vibration signal into a digital signal according to the analog-to-digital conversion module, analyze the digital signal according to the data processing module, and extract the physiological signal corresponding to the user from the digital signal. The physiological signal can be filtered by filtering to separate the heartbeat signal to obtain the heartbeat signal to be detected, and the heartbeat signal to be detected can be divided into a number of heartbeat unit signal segments to be detected to obtain a number of heartbeat unit signal segments to be detected corresponding to the heartbeat signal to be detected.

[0035] The extraction device obtains a quality assessment model, wherein the quality assessment model includes a feature extraction module, a feature processing module and a quality assessment module.

[0036] The ballistocardiogram signal is used in aspects such as heart rate detection, heart rate variability monitoring, cardiac contractility and cardiac output change monitoring.

[0037] The piezoelectric sensor may be a piezoelectric ceramic sensor, a piezoelectric film sensor, etc., and may be placed below the heart in a lying position or under a pillow to obtain the user's human body micro-vibration signal.

[0038] The analog-to-digital conversion module may use an external chip or a corresponding built-in analog-to-digital conversion interface to convert the human body micro-vibration signal into a digital signal.

[0039] The data processing module may use a DSP (Digital Signal Processing) or an ARM (Advanced RISC Machines) processor to analyze the digital signal and extract the user's physiological signal from the digital signal.

[0040] The ballistocardiogram signal is an extremely weak physiological signal with poor stability. Therefore, in the actual acquisition process, interference caused by various factors will affect its signal quality. For example, baseline drift, power frequency noise and other noise problems and body motion artifact problems that may appear in the ballistocardiogram signal affect the reliability of signal feature extraction. In an optional embodiment, the extraction device preprocesses the ballistocardiogram signal to be detected to obtain the preprocessed ballistocardiogram signal to be detected, wherein the preprocessing step includes noise removal processing and body motion artifact removal processing.

[0041] S2: Inputting a plurality of the to-be-detected ballistocardiogram unit signals into the feature extraction module for feature extraction, and obtaining a plurality of initial feature vectors of the to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector.

[0042] In this embodiment, the extraction device performs feature extraction on several of the ballistocardiogram unit signals to be detected to obtain initial feature vectors of several of the ballistocardiogram unit signals to be detected, wherein the initial feature vector includes a signal similarity feature vector, a signal validity feature vector and a signal time-frequency feature vector, and extracts a series of multi-dimensional, objective and quantifiable feature vectors to evaluate the accuracy of the quality assessment of the ballistocardiogram signal.

[0043] See also Figure 2 , Figure 2 The schematic diagram of S2 in the process of the quality assessment method of the ballistocardiogram signal provided by one embodiment of the present application includes steps S201 to S203, which are specifically as follows:

[0044] S201: using a template matching method to obtain template matching results of a plurality of the ballistocardiogram unit signals to be detected, and obtaining first binarization results of a plurality of the ballistocardiogram unit signals to be detected based on the template matching results of the plurality of the ballistocardiogram unit signals to be detected and a preset first binarization algorithm.

[0045] In this embodiment, the extraction device adopts a template matching method to obtain template matching results of several of the to-be-detected ballistocardiogram unit signals. Specifically, the extraction device obtains the highest peak within each interval length of 760ms in several of the to-be-detected ballistocardiogram unit signals through a findpeak algorithm, cuts several of the to-be-detected ballistocardiogram unit signals into several sub-signals with the peak as the center, takes the average of each sub-signal as a template, and uses the template to perform cross-correlation matching on the corresponding ballistocardiogram unit signals to be detected to obtain template matching results of several of the to-be-detected ballistocardiogram unit signals.

[0046] The extraction device obtains first binarization results of the plurality of the to-be-detected ballistocardiogram unit signals according to the template matching results of the plurality of to-be-detected ballistocardiogram unit signals and a preset first binarization algorithm, wherein the first binarization algorithm is:

[0047] x 11 =Round(template_matching(x orig )+z)

[0048] In the formula, x 11 is the first binarization result, x orig is the ballistocardiogram unit signal to be detected, z is the binarization threshold, Round() is the rounding function, and template_matching() is the template matching function.

[0049] S202: Obtaining first envelope signals of a plurality of the ballistocardiogram unit signals to be detected according to the plurality of the ballistocardiogram unit signals to be detected and a preset envelope signal extraction algorithm.

[0050] The envelope signal extraction algorithm is:

[0051]

[0052] In the formula, x env is the envelope signal, and Hilbert( ) is the Hilbert transform function.

[0053] In this embodiment, the extraction device adopts the Hilbert transform method to obtain the Hilbert transform results of several of the ballistocardiogram unit signals to be detected, and obtains the envelope signals of several of the ballistocardiogram unit signals to be detected based on the several ballistocardiogram unit signals to be detected, the Hilbert transform results of several of the ballistocardiogram unit signals to be detected and a preset envelope signal extraction algorithm.

[0054] S203: According to the envelope signals of the plurality of the ballistocardiogram unit signals to be detected and a preset second binarization algorithm, second binarization results of the plurality of the ballistocardiogram unit signals to be detected are obtained, and the first binarization result and the second binarization result of the same ballistocardiogram unit signal to be detected are spliced ​​to obtain first splicing results of the plurality of the ballistocardiogram unit signals to be detected as the signal similarity feature vector.

[0055] In this embodiment, the extraction device obtains the second binarization results of the plurality of the to-be-detected ballistocardiogram unit signals according to the envelope signals of the plurality of to-be-detected ballistocardiogram unit signals and a preset second binarization algorithm, wherein the second binarization algorithm is:

[0056] x 12 =((x env -mean(x env ))>0)

[0057] In the formula, x 12 is the second binarization result, and mean() is the mean function.

[0058] The extraction device splices the first binarization result and the second binarization result of the same cardiac ballistogram unit signal to be detected to obtain the first splicing processing results of several cardiac ballistogram unit signals to be detected, and the large-scale binarization features representing the signal similarity are used as the signal similarity feature vector.

[0059] See also Figure 3 , Figure 3 The schematic diagram of S2 in the process of the quality assessment method of the ballistocardiogram signal provided by one embodiment of the present application includes steps S211 to S212, which are specifically as follows:

[0060] S211: A peak positioning method is used to obtain peak positioning data of several of the to-be-detected ballistocardiogram unit signals; and based on the position data of the H-peak peak point and the L-peak peak point in the peak positioning data of several of the to-be-detected ballistocardiogram unit signals and a preset third binarization algorithm, a third binarization result of several of the to-be-detected ballistocardiogram unit signals is obtained.

[0061] A typical ballistocardiogram signal includes the HIJKL wave group. The quality of the ballistocardiogram signal depends not only on the similarity of the signal in the large-scale time segment, but also on the proportion of the effective signal in the small-scale time segment, that is, the signal segment corresponding to the HIJKL wave group.

[0062] In this embodiment, the extraction device adopts a peak location method to obtain peak location data of a plurality of the ballistocardiogram unit signals to be detected, wherein the peak location data includes the location data of the peak point of the H peak and the location data of the peak point of the L peak.

[0063] Specifically, the extraction device reverses several of the to-be-detected cardiac ballistogram unit signals, uses the findpeak algorithm to preliminarily search for peaks, sets a length threshold interval of 760ms, obtains the two largest target peaks of several of the to-be-detected cardiac ballistogram unit signals within the length threshold interval, and the distance between the two target peaks does not exceed 200ms, marks the positions of the two target peaks, marks the target peak with an earlier position as H peak, and marks the target peak with an earlier position as J peak, and obtains the position data of the peak point of H peak and the peak point of L peak of several of the to-be-detected cardiac ballistogram unit signals.

[0064] The extraction device obtains the third binarization results of the plurality of the to-be-detected ballistocardiogram unit signals according to the position data of the H peak peak point and the position data of the L peak peak point in the peak location data of the plurality of the to-be-detected ballistocardiogram unit signals and a preset third binarization algorithm, wherein the third binarization algorithm is:

[0065] x 21 =STEP(L i -L H )-STEP(L i -L L )

[0066] In the formula, x 21 is the third binarization result, L i is the position data of the i-th signal sampling point of the ballistocardiogram unit signal to be detected, L H is the position data of the peak-to-peak point of the H of the ballistocardiogram unit signal to be detected, L L is the position data of the L peak-to-peak point of the ballistocardiogram unit signal to be detected, and STEP() is a step function.

[0067] S212: Obtain fourth binarization results of several of the ballistocardiogram unit signals to be detected based on envelope signals of the several ballistocardiogram unit signals to be detected and a preset fourth binarization algorithm, and perform splicing processing on the third binarization result and the fourth binarization result of the same ballistocardiogram unit signal to be detected to obtain second splicing processing results of the several ballistocardiogram unit signals to be detected as the signal validity feature vector.

[0068] For the envelope information of the ballistocardiogram signal, if its effectiveness is to be reflected, the envelope curve of the ballistocardiogram should be relatively smooth without mutations and breaks. In this embodiment, the extraction device obtains the differential signal of the envelope signal and the mean data of the WeChat signal according to the envelope signals of the ballistocardiogram unit signals to be detected and the preset fourth binarization algorithm, and uses the mean data to binarize the differential signal of the envelope signal to obtain the fourth binarization results of the ballistocardiogram unit signals to be detected, wherein the fourth binarization algorithm is:

[0069] x 22 =threshold(diff(x env ),mean(diff(x env )))

[0070] In the formula, x 22is the fourth binarization result, threshold() is the threshold function, and diff() is the differential function.

[0071] The extraction device splices the third binarization result and the fourth binarization result of the same cardiac ballistogram unit signal to be detected to obtain a second splicing processing result of several cardiac ballistogram unit signals to be detected, representing the small-scale binarization features of the signal validity as the signal validity feature vector.

[0072] See also Figure 4 , Figure 4 The schematic diagram of S2 in the process of the method for evaluating the quality of the ballistocardiogram signal provided by one embodiment of the present application includes step S221, which is specifically as follows:

[0073] S221: using a short-time Fourier transform method to obtain time-frequency signals of a plurality of the ballistocardiogram unit signals to be detected, scaling the time-frequency signals, and obtaining scaling results of the plurality of the time-frequency signals as the signal time-frequency feature vectors.

[0074] In this embodiment, the extraction device uses a short-time Fourier transform method to obtain time-frequency signals of a plurality of ballistocardiogram unit signals to be detected.

[0075] In order to improve the accuracy of signal quality assessment, the extraction device scales the time-frequency signal to obtain several scale division results of the time-frequency signal as the signal time-frequency feature vector, wherein the scale division results include a first-scale time-frequency signal and a second-scale time-frequency signal.

[0076] Specifically, the extraction device performs noise reduction on several of the time-frequency signals to obtain several time-frequency signals after noise processing as first-scale time-frequency signals of several of the ballistocardiogram unit signals to be detected, wherein the first-scale time-frequency signals are used to reflect the wave group rhythm information of the signal.

[0077] The extraction device divides the time-frequency signals into a plurality of sub-time-frequency signals according to a preset small-scale threshold value as second-scale time-frequency signals of the plurality of ballistocardiogram unit signals to be detected, wherein the second-scale time-frequency signals are used to reflect the wave group detail information of the signal.

[0078] S3: Inputting the initial feature vectors of the plurality of ballistocardiogram unit signals to be detected into the feature processing module for feature processing to obtain the hierarchical feature vectors of the plurality of ballistocardiogram unit signals to be detected.

[0079] In this embodiment, the extraction device inputs the initial feature vectors of several of the cardiac ballistograph unit signals to be detected into the feature processing module for feature processing to obtain hierarchical feature vectors of several of the cardiac ballistograph unit signals to be detected, wherein the hierarchical feature vectors include shallow feature vectors and deep feature vectors.

[0080] See also Figure 5 , Figure 5 The schematic diagram of S3 in the process of the quality assessment method of the ballistocardiogram signal provided by one embodiment of the present application includes steps S31 to S32, which are specifically as follows:

[0081] S31: Obtain an input feature vector by calculating the Hadamard product based on the signal similarity feature vector and the signal validity feature vector of the same ballistocardiogram unit signal to be detected, and obtain shallow feature vectors of several ballistocardiogram unit signals to be detected based on the input feature vector and a preset shallow feature extraction algorithm.

[0082] The shallow feature extraction algorithm is:

[0083]

[0084] In the formula, y wide is the shallow feature vector, w wide is the weight parameter of the shallow feature extraction algorithm, T is the transposition symbol, x in is the input feature vector, b wide is the bias parameter of the shallow feature extraction algorithm.

[0085] In this embodiment, the extraction device calculates the Hadamard product based on the signal similarity feature vector and the signal validity feature vector of the same cardiac ballistogram unit signal to be detected to obtain an input feature vector, and obtains the shallow feature vectors of several cardiac ballistogram unit signals to be detected based on the first input feature vector and a preset shallow feature extraction algorithm.

[0086] S32: obtaining deep feature vectors of a plurality of the ballistocardiogram unit signals to be detected according to the first scale time-frequency signal, the second scale time-frequency signal and a preset deep feature extraction algorithm in the scale division result of the same ballistocardiogram unit signal to be detected.

[0087] The deep feature extraction algorithm is:

[0088] y deep =Convolotion(x dss +resnet_gru(x dsl ))

[0089] In the formula, ydeep is the deep feature vector, x dss is the second scale time-frequency signal, x dsl is the first-scale time-frequency signal, Convolotion() is the convolution function, and resnet_gru() is the residual recurrent network function.

[0090] In this embodiment, the extraction device obtains several deep feature vectors of the ballistocardiogram unit signals to be detected based on the first scale time-frequency signal, the second scale time-frequency signal and the preset deep feature extraction algorithm in the scale division result of the same ballistocardiogram unit signal to be detected.

[0091] S4: Inputting the shallow feature vectors and deep feature vectors of the several ballistocardiogram unit signals to be detected into the quality assessment module to obtain quality assessment results of the several ballistocardiogram unit signals to be detected, and extracting several target ballistocardiogram unit signals from the several ballistocardiogram unit signals to be detected based on the quality assessment results.

[0092] In this embodiment, the extraction device inputs the shallow feature vectors and deep feature vectors of the plurality of ballistocardiogram unit signals to be detected into the quality assessment module to obtain the quality assessment results of the plurality of ballistocardiogram unit signals to be detected.

[0093] See also Figure 6 , Figure 6 The schematic diagram of S4 in the process of the quality assessment method of the ballistocardiogram signal provided by one embodiment of the present application includes steps S41 to S42, which are specifically as follows:

[0094] S41: obtaining predicted probability vectors of a plurality of ballistocardiogram unit signals to be detected based on a shallow feature vector and a deep feature vector of the same ballistocardiogram unit signal to be detected and a preset probability vector calculation algorithm.

[0095] In this embodiment, the extraction device obtains the predicted probability vectors of several ballistocardiogram unit signals to be detected according to the shallow feature vector and the deep feature vector of the same ballistocardiogram unit signal to be detected and a preset probability vector calculation algorithm, wherein the probability vector calculation algorithm is:

[0096] Y = softmax(y wide +w estimate y deep +b estimate )

[0097] Where Y is the predicted probability vector, w estimate is the weight parameter of the quality assessment module, b estimateis the weight parameter of the quality assessment module, and softmax() is a normalized exponential function.

[0098] The extraction device obtains the quality level corresponding to the category with the maximum prediction probability according to the prediction probabilities corresponding to several categories in the prediction probability vector as the quality evaluation result, and obtains the quality evaluation results of several of the to-be-detected ballistocardiogram unit signals.

[0099] S42: According to the prediction probabilities corresponding to several categories in the prediction probability vector, the quality level corresponding to the category with the maximum prediction probability is obtained as the quality evaluation result, and the quality evaluation results of several ballistocardiogram unit signals to be detected are obtained.

[0100] In this embodiment, the extraction device obtains the quality level corresponding to the category with the maximum prediction probability based on the prediction probabilities corresponding to several categories in the prediction probability vector, and obtains the quality evaluation results of several heart ballistograph unit signals to be detected as the quality evaluation result, wherein the quality evaluation results include three categories: A, B, and C.

[0101] According to the quality assessment result, the extraction device uses the to-be-detected ballistocardiogram unit signal segment with a quality assessment result of Class A as the target ballistocardiogram unit signal segment, and extracts a plurality of target ballistocardiogram unit signal segments from the to-be-detected ballistocardiogram unit signal segments. By analyzing the characteristics of the ballistocardiogram signal, extracting the initial feature vector of the ballistocardiogram signal, and performing shallow and deep information mining on the initial feature vector, the direct features and potential features of the ballistocardiogram signal that can affect the quality assessment are obtained, the accuracy of the signal quality assessment result is improved, and thus the accuracy of the extraction of the ballistocardiogram signal is improved.

[0102] In an optional embodiment, the step S5 is also included: training the quality assessment model. Figure 7 , Figure 7 A schematic diagram of S5 in the process of a method for evaluating the quality of a ballistocardiogram signal provided by another embodiment of the present application includes step S51, which is specifically as follows:

[0103] S51: Obtain label data and prediction probability vectors of several sample ballistocardiogram unit signals, and train the quality assessment model according to the label vectors of several categories in the label data of the several sample ballistocardiogram unit signals, the prediction probabilities corresponding to several categories in the prediction probability vectors, and a preset loss function.

[0104] The loss function is:

[0105]

[0106] Where Loss is the loss function, M is the number of sample ballistocardiogram unit signals, C is the number of quality levels, and Y′ ij is the label vector of the jth category of the i-th sample ballistocardiogram unit signal, p ij is the predicted probability of the jth category of the i-th sample ballistocardiogram unit signal.

[0107] In this embodiment, the extraction device obtains the label data and prediction probability vector of several sample cardiocardiogram unit signals, and trains the quality assessment model according to the label vectors of several categories in the label data of several sample cardiocardiogram unit signals, the prediction probabilities corresponding to several categories in the prediction probability vector, and the preset loss function. The model training is carried out in combination with the shallow feature vector and the deep sign vector, which can improve the generalization ability of the model, learn the wave group details of the cardiocardiogram unit signal and the wave group and rhythm information between each cardiocardiogram unit signal, fully consider the objective factors affecting the quality of the cardiocardiogram signal, output a more accurate signal quality assessment result, and improve the accuracy of the extraction of the cardiocardiogram signal.

[0108] Please refer to Figure 8 , Figure 8 This is a schematic diagram of the structure of a device for evaluating the quality of a ballistocardiogram signal provided by an embodiment of the present application. The device can implement all or part of the device for evaluating the quality of a ballistocardiogram signal by software, hardware, or a combination of both. The device 8 includes:

[0109] The data acquisition module 81 is used to obtain the to-be-detected ballistocardiogram signal and a preset quality assessment model, divide the to-be-detected ballistocardiogram signal into a plurality of to-be-detected ballistocardiogram unit signals, and obtain a plurality of to-be-detected ballistocardiogram unit signals corresponding to the to-be-detected ballistocardiogram signal, wherein the quality assessment model includes a feature extraction module, a feature processing module, and a quality assessment module;

[0110] An initial feature extraction module 82 is used to input a plurality of the to-be-detected ballistocardiogram unit signals into the feature extraction module for feature extraction, and obtain initial feature vectors of the plurality of to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector;

[0111] A hierarchical feature extraction module 83 is used to input the initial feature vectors of the plurality of ballistocardiogram unit signals to be detected into the feature processing module for feature processing to obtain the hierarchical feature vectors of the plurality of ballistocardiogram unit signals to be detected;

[0112] The signal quality assessment module 84 is used to input the hierarchical feature vectors of the several ballistocardiogram unit signals to be detected into the quality assessment module, obtain the quality assessment results of the several ballistocardiogram unit signals to be detected, and extract several target ballistocardiogram unit signals from the several ballistocardiogram unit signals to be detected based on the quality assessment results.

[0113] In an embodiment of the present application, a to-be-detected ballistocardiogram signal and a preset quality assessment model are obtained through a data acquisition module, the to-be-detected ballistocardiogram signal is divided into a plurality of to-be-detected ballistocardiogram unit signals, and a plurality of to-be-detected ballistocardiogram unit signals corresponding to the to-be-detected ballistocardiogram signal are obtained, wherein the quality assessment model includes a feature extraction module, a feature processing module, and a quality assessment module; through an initial feature extraction module, feature extraction is performed on a plurality of to-be-detected ballistocardiogram unit signals, and initial feature vectors of a plurality of to-be-detected ballistocardiogram unit signals are obtained, wherein the initial feature vector includes a signal similarity feature vector, a signal validity feature vector, and a signal The invention relates to a method for extracting time-frequency feature vectors of a plurality of ballistocardiogram unit signals to be detected; inputting the initial feature vectors of a plurality of ballistocardiogram unit signals to be detected into the feature processing module through a hierarchical feature extraction module for feature processing, and obtaining hierarchical feature vectors of a plurality of ballistocardiogram unit signals to be detected, wherein the feature vectors include shallow feature vectors and deep feature vectors; inputting the hierarchical feature vectors of a plurality of ballistocardiogram unit signals to be detected into the quality assessment module through a signal quality assessment module, and obtaining quality assessment results of a plurality of ballistocardiogram unit signals to be detected; and extracting a plurality of target ballistocardiogram unit signals from a plurality of ballistocardiogram unit signals to be detected according to the quality assessment results. By analyzing the features of the ballistocardiogram signal, extracting the initial feature vector of the ballistocardiogram signal, and performing shallow and deep information mining on the initial feature vector, the direct features and potential features of the ballistocardiogram signal that can affect the quality assessment are obtained, and the accuracy of the signal quality assessment results is improved, thereby improving the accuracy of the extraction of the ballistocardiogram signal.

[0114] Please refer to Fig. 9 , Fig. 9 The computer device 9 is a schematic diagram of a structure of a computer device provided in an embodiment of the present application. The computer device 9 includes: a processor 91, a memory 92, and a computer program 93 stored in the memory 92 and executable on the processor 91; the computer device may store a plurality of instructions, which are suitable for being loaded and executed by the processor 91 Figures 1 to 7 The method steps of the embodiment shown in the figure can be found in the specific implementation process. Figures 1 to 7 The specific description of the illustrated embodiment will not be repeated here.

[0115] Among them, the processor 91 may include one or more processing cores. The processor 91 uses various interfaces and lines to connect various parts in the server, and executes various functions and processes data of the quality assessment device 8 of the ballistocardiogram signal by running or executing instructions, programs, code sets or instruction sets stored in the memory 92, and calling the data in the memory 92. Optionally, the processor 91 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programble Logic Array, PLA). The processor 91 can integrate one or more combinations of a central processing unit 91 (Central Processing Unit, CPU), an image processor 91 (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the touch display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 91, and it can be implemented by a single chip.

[0116] Among them, the memory 92 may include a random access memory 92 (Random Access Memory, RAM), and may also include a read-only memory 92 (Read-Only Memory). Optionally, the memory 92 includes a non-transitory computer-readable storage medium. The memory 92 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 92 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch instructions, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 92 may also be optionally at least one storage device located away from the aforementioned processor 91.

[0117] The present application also provides a storage medium that can store multiple instructions, which are suitable for the processor to load and execute the above-mentioned Figures 1 to 7 For the method steps of the embodiment, the specific execution process can be found in Figures 1 to 7 The specific description of the embodiment will not be repeated here.

[0118] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0119] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0120] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraint algorithm of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0121] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0122] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0123] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0124] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc.

[0125] The present invention is not limited to the above-mentioned embodiments. If various changes or modifications to the present invention do not depart from the spirit and scope of the present invention, and if these changes and modifications fall within the scope of the claims and equivalent technologies of the present invention, the present invention is also intended to include these changes and modifications.

Claims

1. A method for evaluating the quality of a ballistocardiogram signal, characterized in that: The following steps are involved: Obtaining a ballistocardiogram signal to be detected and a preset quality assessment model, dividing the ballistocardiogram signal to be detected into a plurality of ballistocardiogram unit signals to be detected, and obtaining a plurality of ballistocardiogram unit signals to be detected corresponding to the ballistocardiogram signal to be detected, wherein the quality assessment model includes a feature extraction module, a feature processing module and a quality assessment module; Inputting a plurality of the to-be-detected ballistocardiogram unit signals into the feature extraction module for feature extraction, and obtaining initial feature vectors of the plurality of to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector; A template matching method is adopted to obtain template matching results of a plurality of the ballistocardiogram unit signals to be detected, and a first binarization result of a plurality of the ballistocardiogram unit signals to be detected is obtained according to the template matching results of the plurality of the ballistocardiogram unit signals to be detected and a preset first binarization algorithm, wherein the first binarization algorithm is: In the formula, is the first binarization result, is the ballistocardiogram unit signal to be detected, is the binarization threshold, is the rounding function, is the template matching function; According to the plurality of the to-be-detected ballistocardiogram unit signals and a preset envelope signal extraction algorithm, envelope signals of the plurality of to-be-detected ballistocardiogram unit signals are obtained, wherein the envelope signal extraction algorithm is: In the formula, is the envelope signal, is the Hilbert transform function; According to the envelope signals of the plurality of the to-be-detected ballistocardiogram unit signals and a preset second binarization algorithm, second binarization results of the plurality of the to-be-detected ballistocardiogram unit signals are obtained, and the first binarization result and the second binarization result of the same to-be-detected ballistocardiogram unit signal are spliced ​​to obtain first splicing results of the plurality of to-be-detected ballistocardiogram unit signals as the signal similarity feature vector; A peak location method is used to obtain peak location data of a plurality of the to-be-detected ballistocardiogram unit signals, and a peak location data of a plurality of the to-be-detected ballistocardiogram unit signals is obtained. H Peak-to-peak point location data, L The position data of the peak-to-peak point and the preset third binarization algorithm are used to obtain the third binarization results of the plurality of the ballistocardiogram unit signals to be detected, wherein the third binarization algorithm is: In the formula, is the third binarization result, is the first i The position data of the signal sampling points, is the ballistocardiogram unit signal to be detected H The peak-to-peak point location data, is the ballistocardiogram unit signal to be detected L The peak-to-peak point location data, is a step function; According to the envelope signals of the plurality of the to-be-detected ballistocardiogram unit signals and the preset fourth binarization algorithm, the fourth binarization results of the plurality of the to-be-detected ballistocardiogram unit signals are obtained, and the third binarization result and the fourth binarization result of the same to-be-detected ballistocardiogram unit signal are spliced ​​to obtain the second splicing results of the plurality of the to-be-detected ballistocardiogram unit signals as the signal validity feature vector, wherein the fourth binarization algorithm is: In the formula, is the fourth binarization result, is the threshold function, is a differential function; Adopting a short-time Fourier transform method to obtain time-frequency signals of a plurality of the ballistocardiogram unit signals to be detected, performing scale division on the time-frequency signals, and obtaining scale division results of the plurality of the time-frequency signals as the signal time-frequency feature vectors, wherein the scale division results include a first-scale time-frequency signal and a second-scale time-frequency signal, the first-scale time-frequency signal is used to reflect the wave group rhythm information of the signal, and the second-scale time-frequency signal is used to reflect the wave group detail information of the signal; Inputting the initial feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals into the feature processing module for feature processing, and obtaining the hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals, wherein the hierarchical feature vectors include shallow feature vectors and deep feature vectors; The Hadamard product is calculated based on the signal similarity feature vector and the signal validity feature vector of the same ballistocardiogram unit signal to be detected to obtain an input feature vector, and the shallow feature vectors of several ballistocardiogram unit signals to be detected are obtained based on the input feature vector and a preset shallow feature extraction algorithm, wherein the shallow feature extraction algorithm is: In the formula, is the shallow feature vector, is the weight parameter of the shallow feature extraction algorithm, T is the transpose symbol, is the input feature vector, is the bias parameter of the shallow feature extraction algorithm; According to the first scale time-frequency signal, the second scale time-frequency signal and the preset deep feature extraction algorithm in the scale division result of the same ballistocardiogram unit signal to be detected, a plurality of deep feature vectors of the ballistocardiogram unit signal to be detected are obtained, wherein the deep feature extraction algorithm is: In the formula, is the deep feature vector, is the second scale time-frequency signal, is the first scale time-frequency signal, is the convolution function, is the residual recurrent network function; The hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals are input into the quality assessment module to obtain quality assessment results of the plurality of the to-be-detected ballistocardiogram unit signals, and based on the quality assessment results, a plurality of target ballistocardiogram unit signals are extracted from the plurality of the to-be-detected ballistocardiogram unit signals.

2. The method for evaluating the quality of a ballistocardiogram signal according to claim 1, wherein: The method of inputting shallow feature vectors and deep feature vectors of a plurality of the to-be-detected ballistocardiogram unit signals into the quality assessment module to obtain quality assessment results of a plurality of the to-be-detected ballistocardiogram unit signals comprises the following steps: According to the shallow feature vector and the deep feature vector of the same ballistocardiogram unit signal to be detected and a preset probability vector calculation algorithm, a plurality of predicted probability vectors of the ballistocardiogram unit signals to be detected are obtained; according to the predicted probabilities corresponding to the plurality of categories in the predicted probability vector, a quality grade corresponding to the category with the maximum predicted probability is obtained as the quality evaluation result, and a plurality of quality evaluation results of the ballistocardiogram unit signals to be detected are obtained, wherein the probability vector calculation algorithm is: In the formula, is the predicted probability vector, is the weight parameter of the quality assessment module, is the bias parameter of the quality assessment module, is the normalized exponential function.

3. The method for evaluating the quality of a ballistocardiogram signal according to claim 2, wherein: The method further comprises the steps of: training the quality assessment model, wherein the training of the quality assessment model comprises the steps of: Obtain label data and prediction probability vectors of several sample ballistocardiogram unit signals, and train the quality assessment model according to label vectors of several categories in the label data of several sample ballistocardiogram unit signals, prediction probabilities corresponding to several categories in the prediction probability vectors, and a preset loss function, wherein the loss function is: In the formula, is the loss function, M is the number of sample ballistocardiogram unit signals, C is the number of quality levels, For the i The first sample ballistocardiogram unit signal j The label vector of each category, For the i The first sample ballistocardiogram unit signal j The predicted probability of each class.

4. A device for the method for evaluating the quality of a ballistocardiogram signal according to claim 1, characterized in that: include: A data acquisition module, used for obtaining a to-be-detected ballistocardiogram signal and a preset quality assessment model, dividing the to-be-detected ballistocardiogram signal into a plurality of to-be-detected ballistocardiogram unit signals, and obtaining a plurality of to-be-detected ballistocardiogram unit signals corresponding to the to-be-detected ballistocardiogram signal, wherein the quality assessment model comprises a feature extraction module, a feature processing module and a quality assessment module, and the feature processing module comprises a shallow feature processing module and a deep feature processing module; An initial feature extraction module, used for extracting features from a plurality of the to-be-detected ballistocardiogram unit signals, and obtaining initial feature vectors of the plurality of to-be-detected ballistocardiogram unit signals, wherein the initial feature vectors include a signal similarity feature vector, a signal validity feature vector, and a signal time-frequency feature vector; A hierarchical feature extraction module, used for inputting the initial feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals into the feature processing module for feature processing, and obtaining the hierarchical feature vectors of the plurality of the to-be-detected ballistocardiogram unit signals, wherein the hierarchical feature vectors include shallow feature vectors and deep feature vectors; A signal quality assessment module is used to input the hierarchical feature vectors of the several ballistocardiogram unit signals to be detected into the quality assessment module, obtain quality assessment results of the several ballistocardiogram unit signals to be detected, and extract several target ballistocardiogram unit signals from the several ballistocardiogram unit signals to be detected based on the quality assessment results.

5. A computer device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for evaluating the quality of a ballistocardiogram signal as claimed in any one of claims 1 to 3 when executing the computer program.

6. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for evaluating the quality of a ballistocardiogram signal according to any one of claims 1 to 3 are implemented.

Images