CN116965812A - Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis - Google Patents
Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis Download PDFInfo
- Publication number
- CN116965812A CN116965812A CN202311001051.1A CN202311001051A CN116965812A CN 116965812 A CN116965812 A CN 116965812A CN 202311001051 A CN202311001051 A CN 202311001051A CN 116965812 A CN116965812 A CN 116965812A
- Authority
- CN
- China
- Prior art keywords
- blood glucose
- fourier transform
- blood
- blood flow
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000008280 blood Substances 0.000 title claims abstract description 88
- 210000004369 blood Anatomy 0.000 title claims abstract description 88
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 title claims abstract description 55
- 239000008103 glucose Substances 0.000 title claims abstract description 55
- 238000004458 analytical method Methods 0.000 title claims abstract description 31
- 238000001514 detection method Methods 0.000 title claims abstract description 23
- 230000017531 blood circulation Effects 0.000 claims abstract description 42
- 238000000034 method Methods 0.000 claims abstract description 40
- 238000007637 random forest analysis Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000001228 spectrum Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 8
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 32
- 230000008569 process Effects 0.000 claims description 19
- 238000000354 decomposition reaction Methods 0.000 claims description 14
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000010183 spectrum analysis Methods 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 3
- 206010012601 diabetes mellitus Diseases 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 235000005911 diet Nutrition 0.000 description 2
- 230000037213 diet Effects 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000000691 measurement method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 244000052616 bacterial pathogen Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002512 chemotherapy Methods 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000000276 sedentary effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
- A61B5/1455—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue using optical sensors, e.g. spectral photometrical oximeters
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
- A61B5/0075—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence by spectroscopy, i.e. measuring spectra, e.g. Raman spectroscopy, infrared absorption spectroscopy
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/026—Measuring blood flow
- A61B5/0261—Measuring blood flow using optical means, e.g. infrared light
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/026—Measuring blood flow
- A61B5/0295—Measuring blood flow using plethysmography, i.e. measuring the variations in the volume of a body part as modified by the circulation of blood therethrough, e.g. impedance plethysmography
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
- A61B5/14532—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring glucose, e.g. by tissue impedance measurement
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/40—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for data related to laboratory analysis, e.g. patient specimen analysis
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Heart & Thoracic Surgery (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Physiology (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Hematology (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Cardiology (AREA)
- Optics & Photonics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Emergency Medicine (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The application discloses a noninvasive blood glucose detection method and system based on fractional Fourier transform analysis, comprising the following steps: collecting blood flow signals of a human body through a PPG signal sensor, and measuring fingertip blood glucose concentration through a blood glucose meter at the same time, so that the blood flow signals correspond to the fingertip blood glucose concentration; preprocessing a blood flow signal, performing fractional discrete Fourier transform analysis, extracting an energy spectrum as a data characteristic, and constructing a data set according to fingertip blood sugar concentration; training through a data set based on a random forest regression model, and constructing a blood sugar prediction model after evaluating model prediction errors and accuracy of predicting blood sugar values; acquiring fingertip blood glucose concentration of a target to be detected by acquiring a blood flow signal of the target to be detected based on a blood glucose prediction model; the method and the device obtain the dynamic characteristics between the time domain and the frequency domain of the signal through fractional Fourier transform, well overcome the defect of classical Fourier transform, and furthest reserve useful information contained in the original data.
Description
Technical Field
The application relates to the technical field of blood glucose prediction, in particular to a noninvasive blood glucose detection method and system based on fractional Fourier transform analysis.
Background
In recent years, the sedentary lifestyle has become a daily routine for people due to rapid economic development, urbanization and nutritional transformation. The number of diabetic and obese patients is greatly increasing, both in developed and developing countries. While the main burden is now in developing countries, about 80% of diabetics live in low and medium income countries and communities. The average age of diabetics also gradually decreases. Diabetes is a global epidemic that is highly ranked on the international health agenda as a threat to human health and global economy. Depending on the current medical level, we cannot cure diabetes completely, but only rely on drugs or chemotherapy to inhibit the condition from further worsening. By knowing the medical history of a diabetic patient and predicting the likely outcome in the future, it will help the physician to better understand the condition of the patient and provide better, higher quality treatment. Therefore, the intervention of diabetics with diseases is particularly important. Studies have shown that if diabetics adhere to proper diet and medication every day, diabetes will be effectively controlled, and continuous blood glucose monitoring is required, diabetics can adjust the three meal diet intake according to their own blood glucose profile. The blood sugar detection at the present stage is mainly invasive detection, but the invasive detection not only needs to use blood taking needles and test paper for each detection, thereby causing environmental pollution, but also needs to extract blood from fingertips for each detection, so that the wound of a patient is difficult to heal in time, and the risk of causing wound to infect germs is generated. Although the results of invasive tests are accurate and can be an important basis for diagnosing diabetes, the method is not suitable for continuous monitoring of diabetics, both from the point of view of the patient and the medical device.
Based on various drawbacks of invasive blood glucose monitoring, most diabetics desire blood glucose monitoring by non-invasive methods. The prior art discloses a time-frequency domain comprehensive analysis noninvasive blood glucose measurement method, which comprises the steps of firstly obtaining a PPG signal of a fingertip, denoising the PPG signal, analyzing the signal by using fast Fourier transform, extracting meaningful time domain and frequency domain characteristics, and finally predicting a blood glucose value by establishing a random forest model to realize noninvasive blood glucose prediction. But PPG signals are mostly non-stationary signals, exposing the limitations of classical fourier transforms. For the nonstationary signal, only the time domain or frequency domain features are selected to predict the blood sugar, so that a great number of effective features in the middle of the time domain and the frequency domain are inevitably lost, and the blood sugar prediction result is not accurate enough, therefore, a noninvasive blood sugar detection method and system based on fractional Fourier transform analysis are urgently needed, and the technical problems are solved.
Disclosure of Invention
In order to solve the problem of loss of effective features, the application provides a noninvasive blood glucose measurement method and a noninvasive blood glucose measurement system based on discrete fractional Fourier transform.
In order to achieve the technical aim, the application provides a noninvasive blood glucose detection method based on fractional Fourier transform analysis, which comprises the following steps:
collecting blood flow signals of a human body through a PPG signal sensor, and measuring fingertip blood glucose concentration through a blood glucose meter at the same time, so that the blood flow signals correspond to the fingertip blood glucose concentration;
preprocessing a blood flow signal, performing fractional discrete Fourier transform analysis, extracting an energy spectrum as a data characteristic, and constructing a data set according to fingertip blood sugar concentration;
training through a data set based on a random forest regression model, and constructing a blood sugar prediction model after evaluating model prediction errors and accuracy of predicting blood sugar values;
based on the blood glucose prediction model, the fingertip blood glucose concentration of the target to be measured is obtained by collecting the blood flow signal of the target to be measured.
Preferably, in the process of preprocessing the blood flow signal, the influence caused by baseline drift is filtered through empirical mode decomposition, and a singular spectrum analysis method is used for removing noise, so that the main components of the original signal are left.
Preferably, in the process of performing empirical mode decomposition on the blood flow signal, the blood flow signal is decomposed into a finite number of eigenmode functions by empirical mode decomposition, and each of the decomposed IMF components includes local feature signals of different time scales of the original signal.
Preferably, in the process of performing empirical mode decomposition on the blood flow signal, fitting an upper envelope curve of the blood flow signal by using cubic spline interpolation according to a maximum point of a first signal sequence of the blood flow signal;
fitting a lower envelope curve of the blood flow signal by using cubic spline interpolation according to the minimum value point of the signal sequence;
obtaining the average value of the upper envelope curve and the lower envelope curve, subtracting the average value from the first signal sequence, and obtaining a second signal sequence for representing the eigenmode function, wherein the eigenmode function meets the following conditions:
condition 1: the number of the local extreme points and the number of the zero crossing points of the function are equal or differ by one in the whole time range;
condition 2: at any point in time, the average of the upper envelope and the lower envelope is less than the set threshold.
Preferably, in the process of acquiring the second signal sequence, the first signal sequence subtracts the mean value to generate a third signal sequence which is not an eigenmode function, then the third signal sequence is defined as a new signal sequence, and an upper envelope line, a lower envelope line and a mean value of the upper envelope line and the lower envelope line of the third signal sequence are acquired until the second signal sequence is acquired, wherein a residual error is defined according to the second signal sequence, and if the residual error is an eigenmode function or a monotonic function, the residual error indicates that the empirical mode decomposition of the blood flow signal is completed.
Preferably, in the process of performing fractional discrete fourier transform analysis on the preprocessed blood flow signal, the process of fractional discrete fourier transform analysis is as follows:
step S1: let transform order p=0;
step S2: performing p-order discrete fractional Fourier transform on the second signal sequence to obtain frequency domain information and frequency domain characteristic parameters, and extracting a spectrum set in the middle energy set as a data characteristic;
step S3: let p=p+0.1, repeat step S2 until p=1.
Step S4: the 11 data features are spliced into a vector which is used as a total feature vector, and a data set is constructed according to the fingertip blood glucose concentration corresponding to the blood flow signal.
Preferably, in the process of constructing the blood glucose prediction model, based on a random forest regression model, the data set is divided into 5 equal parts after being randomly disturbed, five-fold cross validation is used, and the data set is classified and regressed according to the average value of the 5 obtained results as a final regression result, so that the blood glucose prediction model for noninvasive detection of blood glucose is constructed.
The application also discloses a noninvasive blood glucose detection system based on fractional Fourier transform analysis, which comprises:
the data acquisition module is used for acquiring blood flow signals of a human body through the PPG signal sensor, and measuring fingertip blood sugar concentration through the blood glucose meter so that the blood flow signals correspond to the fingertip blood sugar concentration;
the data processing module is used for preprocessing the blood flow signals, performing fractional discrete Fourier transform analysis, extracting an energy spectrum as a data characteristic, and constructing a data set according to the fingertip blood sugar concentration;
the blood sugar detection module is used for training through a data set based on a random forest regression model, constructing a blood sugar prediction model after evaluating model prediction errors and accuracy of predicting blood sugar values, and acquiring fingertip blood sugar concentration of a target to be detected by collecting blood flow signals of the target to be detected based on the blood sugar prediction model.
The application discloses the following technical effects:
the method and the device obtain the dynamic characteristics between the time domain and the frequency domain of the signal through fractional Fourier transform, well overcome the defect of classical Fourier transform, and furthest reserve useful information contained in the original data.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a process of a non-invasive glucose monitoring method based on discrete fractional Fourier transform analysis according to an embodiment of the present application;
fig. 2 is a rotation process of the fractional fourier transform on the time-frequency plane according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. The components of the embodiments of the present application generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the application, as presented in the figures, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. All other embodiments, which can be made by a person skilled in the art without making any inventive effort, are intended to be within the scope of the present application.
As shown in fig. 1-2, example 1: as shown in fig. 1, the application provides a noninvasive blood glucose monitoring method for discrete fractional fourier transform analysis, which comprises the following steps:
step 1, the ppg signal sensor can continuously and accurately measure weak blood flow signals. The PPG signal sensor is first used to collect blood flow signals of the human body while the fingertip blood glucose concentration is measured with the portable blood glucose meter to be used as a reference value for the following prediction process.
And 2, preprocessing the collected PPG signals, removing noise and the like, wherein the preprocessing mainly uses Empirical Mode Decomposition (EMD) to filter the influence caused by baseline drift, and then uses a singular spectrum analysis method to remove noise, so that the main components of the original signals are left.
And step 3, carrying out fractional discrete Fourier transform analysis on the obtained signals, and extracting the concentrated energy spectrum as data characteristics.
And 4, selecting a random forest regression model to predict the blood sugar value, comparing the blood sugar value with the originally measured blood sugar result by using five-fold cross verification, and evaluating the model prediction error.
And 5, predicting the accuracy of the blood glucose value by using a Clark error grid analysis.
The empirical mode decomposition used in step 2 enables the complex signal to be decomposed into a finite number of eigen-mode functions (IMFs), each IMF component being decomposed to contain local feature signals of different time scales of the original signal. The method comprises the following steps:
step S1: giving a signal sequence x (t), finding all maximum points of the x (t), and fitting an upper envelope curve of the original signal by using cubic spline interpolation; similarly, all minimum value points of x (t) are found, and a lower envelope curve of the original data is fitted by using cubic spline interpolation.
Step S2: calculating the average value of the upper envelope curve and the lower envelope curve, and marking the average value as m;
step S3: subtracting the average envelope m from the original signal sequence to obtain a new signal sequence h (t), namely: x (t) -m=h (t);
step S4: judging whether h (t) is an eigenmode function, and one eigenmode function must satisfy the following two conditions: the method includes the steps that the number of local extreme points and zero crossing points of a function in the whole time range must be equal or at most differ by one; the envelope of the local maxima (upper envelope) and the envelope of the local minima (lower envelope) must be less than the set threshold value on average at any point in time.
Step S5: if h (t) is not an eigenmode function, defining h (t) as a new signal sequence, repeating the steps S1-4, if h (t) is an eigenmode function, defining a residual error r (t) =x (t) -h (t), and if r (t) is an eigenfunction or a monotonic function, completing decomposition; otherwise, defining the residual error r (t) as a new signal sequence, and repeating the steps S1-4.
In the step 2, singular spectrum analysis is adopted as a denoising method, and the method specifically comprises the following steps:
step S1: given a signal time sequence x 1 ，x 2 ，...，x N ]N is the sequence length. Firstly, selecting a proper window length L and a proper step length P, and performing hysteresis sequencing on the time sequence to obtain a track matrix:
step S2: singular value decomposition is performed on the trajectory matrix X, i.e., X is decomposed into the following forms:
X＝U∑V T
wherein U and V are unitary matrices, T represents transpose, Σ is a diagonal matrix, and each diagonal element is a singular value of X.
Step S3: arranging the diagonal elements of the sigma in the step S2 in sequence from large to small, wherein the larger the element value is, the more important the element value is, and then the first n main components are selected to form a new time sequence, namely reconstruction is realized.
The definition of the fractional fourier transform in step 3 can be expressed as:
wherein alpha is the rotation angleT, u represent arguments in the kernel function. K (K) α (t, u) is a kernel function of a fractional Fourier transform, defined as
When α=2npi, K α (t，u)＝δ(t-u)；
When α= (2n±1) pi, K α (t，u)＝δ(t+u)；
The fractional fourier transform can also be expressed as:
defining p as the order of the fourier transform, α as the rotation angle, the relationship between the two being:
it can be seen that when the angle α is equal to pi/2, p=1, and the fractional fourier transform is the classical fourier transform.
The fractional fourier transform satisfies some properties, some of which, alpha, are used in the subsequent computation of the discrete fractional fourier transform, as shown in table 1 1 And alpha 2 Are real numbers, and each represents the number of fourier transforms. c n Representation and x n An independent expression. X is x n Representing a function of the fourier transform.
TABLE 1
Fig. 2 shows a rotation process of the fractional fourier transform on a time-frequency plane, and it can be seen that, when the rotation angle α increases from 0 to 90 degrees, the fractional fourier transform can obtain dynamic characteristics between a time domain and a frequency domain of a signal, and is not limited to analyzing the signal on a coordinate axis (time domain or frequency domain), so that the defect of classical fourier transform is well overcome, and an optimal processing result in a global sense is found.
The above-mentioned fractional fourier transform is used for researching continuous signals, but the signals acquired in reality can only be discrete signals, and are difficult to solve by an analytic method, so that we usually use a numerical calculation method, and therefore, the fractional fourier transform needs to be converted into the discrete fractional fourier transform.
Assume that the wigner-ville distribution of signal x (t) is defined within a circle of diameter deltax centered on the origin. If 0.5 < a < 1.5, the signal x (t) is multiplied by a chirp signal and then has a bandwidth deltax in the frequency domain. Expressed by shannon interpolation:
in the step 3, fractional discrete fourier transform analysis is performed on the obtained signal, and since the fractional fourier transform has periodicity and symmetry of coordinate axes, only the transformation of the signal between the first quadrant rotations, that is, the transformation order p e [0,1] needs to be studied. The selection of the orders can analyze the signals according to the need, and the time-frequency signals with more information can be selected at equal intervals according to the selection of the orders, specifically:
step S1: let the order p=0;
step S2: performing p-order discrete fractional Fourier transform on the signal sequence x (t) subjected to pretreatment and denoising to obtain frequency domain information and frequency domain characteristic parameters, extracting a spectrum set in a middle energy comparison set as a data characteristic, and setting the data characteristic as [ x ] 1 ，x 2 ，...，x L ]。
Step S3: let p=p+0.1, repeat step S2 until p=1.
Step S4: the 11 data features are spliced into a vector as a total feature vector.
And 4, selecting a random forest regression model to predict the blood sugar value, wherein the method specifically comprises the following steps:
and (3) putting the total feature vector in the step (3) into a random forest model, constructing a prediction model by sampling the data by the random forest, namely generating a plurality of decision trees, and sequentially classifying and regressing the data. In the process, the data set is divided into 5 equal parts after being randomly disturbed, five-fold cross validation is used, one part is taken for testing in each experiment, and the rest is used for training. Taking the first part as a test set in the first experiment, and taking the rest as a training set; the second experiment takes the second set as the test set, the rest as the training set, and so on. The results obtained from the experiment 5 times were averaged as the final regression result.
In step 4, estimating a model prediction error, wherein specific indexes are a correlation coefficient and a mean absolute relative error (MARD), and a correlation coefficient formula is as follows:
whereas the MARD formula is:
wherein x (t) is a predicted result, y (t) is a true result,is the mean value of the x (t) samples, +.>The mean value of x (t) samples, and N is the number of samples.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In the description of the present application, it should be understood that the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present application, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (8)
1. The noninvasive blood glucose detection method based on fractional Fourier transform analysis is characterized by comprising the following steps of:
collecting blood flow signals of a human body through a PPG signal sensor, and simultaneously measuring fingertip blood sugar concentration through a blood glucose meter, so that the blood flow signals correspond to the fingertip blood sugar concentration;
preprocessing the blood flow signal, performing fractional discrete Fourier transform analysis, extracting an energy spectrum as a data characteristic, and constructing a data set according to the fingertip blood glucose concentration;
training through the data set based on a random forest regression model, and constructing a blood glucose prediction model after evaluating model prediction errors and accuracy of predicting blood glucose values;
and acquiring the fingertip blood glucose concentration of the target to be detected by acquiring a blood flow signal of the target to be detected based on the blood glucose prediction model.
2. The method for noninvasive blood glucose detection based on fractional fourier transform analysis of claim 1, wherein:
in the process of preprocessing the blood flow signals, the influence caused by baseline drift is filtered through empirical mode decomposition, and a singular spectrum analysis method is used for removing noise, so that the main components of the original signals are left.
3. The method for noninvasive blood glucose detection based on fractional fourier transform analysis of claim 2, wherein:
in the process of carrying out empirical mode decomposition on a blood flow signal, the blood flow signal is decomposed into a limited number of eigenmode functions through the empirical mode decomposition, and each decomposed IMF component comprises local characteristic signals of different time scales of the original signal.
4. A method of noninvasive blood glucose detection based on fractional fourier transform analysis as claimed in claim 3, wherein:
in the process of performing empirical mode decomposition on a blood flow signal, fitting an upper envelope curve of the blood flow signal by using cubic spline interpolation according to a maximum point of a first signal sequence of the blood flow signal;
fitting a lower envelope curve of the blood flow signal by using cubic spline interpolation according to the minimum value point of the signal sequence;
obtaining the average value of the upper envelope curve and the lower envelope curve, subtracting the average value from the first signal sequence, and obtaining a second signal sequence for representing an eigenmode function, wherein the eigenmode function meets the following conditions:
condition 1: the number of the local extreme points and the number of the zero crossing points of the function are equal or differ by one in the whole time range;
condition 2: at any point in time, the average of the upper envelope and the lower envelope is less than a set threshold.
5. The method for noninvasive blood glucose detection based on fractional fourier transform analysis of claim 4, wherein:
in the process of obtaining the second signal sequence, the first signal sequence subtracts the mean value to generate a third signal sequence which is not the eigen-mode function, then the third signal sequence is defined as a new signal sequence, and an upper envelope line, a lower envelope line and the mean value of the upper envelope line and the lower envelope line of the third signal sequence are obtained until the second signal sequence is obtained, wherein a residual error is defined according to the second signal sequence, and if the residual error is an eigen-mode function or a monotonic-mode function, the empirical mode decomposition of the blood flow signal is completed.
6. The method for noninvasive blood glucose detection based on fractional fourier transform analysis of claim 5, wherein:
in the process of carrying out fractional discrete Fourier transform analysis on the preprocessed blood flow signal, the process of fractional discrete Fourier transform analysis is as follows:
step S1: let transform order p=0;
step S2: performing p-order discrete fractional Fourier transform on the second signal sequence to acquire frequency domain information and frequency domain characteristic parameters, and extracting a spectrum set in the middle energy set as a data characteristic;
step S3: let p=p+0.1, repeat step S2 until p=1.
Step S4: and splicing the 11 data features into a vector serving as a total feature vector, and constructing the data set according to the fingertip blood glucose concentration corresponding to the blood flow signal.
7. The method for noninvasive blood glucose detection based on fractional fourier transform analysis of claim 6, wherein:
in the process of constructing the blood glucose prediction model, the data set is divided into 5 equal parts after being randomly disturbed based on a random forest regression model, five-fold cross verification is used, and the data set is classified and regressed according to the average value of the 5 obtained results as a final regression result, so that the blood glucose prediction model for noninvasive detection of blood glucose is constructed.
8. A non-invasive blood glucose testing system based on fractional fourier transform analysis, comprising:
the data acquisition module is used for acquiring blood flow signals of a human body through the PPG signal sensor, and measuring fingertip blood sugar concentration through the blood glucose meter, so that the blood flow signals correspond to the fingertip blood sugar concentration;
the data processing module is used for preprocessing the blood flow signals, then carrying out fractional discrete Fourier transform analysis, extracting energy spectrum as data characteristics, and constructing a data set according to the fingertip blood sugar concentration;
the blood sugar detection module is used for training through the data set based on a random forest regression model, constructing a blood sugar prediction model after evaluating model prediction errors and accuracy of predicting blood sugar values, and acquiring fingertip blood sugar concentration of a target to be detected by collecting blood flow signals of the target to be detected based on the blood sugar prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311001051.1A CN116965812A (en) | 2023-08-10 | 2023-08-10 | Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311001051.1A CN116965812A (en) | 2023-08-10 | 2023-08-10 | Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116965812A true CN116965812A (en) | 2023-10-31 |
Family
ID=88484857
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311001051.1A Pending CN116965812A (en) | 2023-08-10 | 2023-08-10 | Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116965812A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104635223A (en) * | 2015-01-08 | 2015-05-20 | 南京大学 | Laser echo denoising method based on empirical mode decomposition and fractional Fourier transformation |
| US20160256114A1 (en) * | 2015-03-02 | 2016-09-08 | Guilin Medicine Electronic Technology Co., Ltd. | Non-invasive blood sugar measuring method and fingertip measuring probe |
| CN110338813A (en) * | 2019-06-04 | 2019-10-18 | 西安理工大学 | A kind of Noninvasive Blood Glucose Detection Methods based on spectrum analysis |
-
2023
- 2023-08-10 CN CN202311001051.1A patent/CN116965812A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104635223A (en) * | 2015-01-08 | 2015-05-20 | 南京大学 | Laser echo denoising method based on empirical mode decomposition and fractional Fourier transformation |
| US20160256114A1 (en) * | 2015-03-02 | 2016-09-08 | Guilin Medicine Electronic Technology Co., Ltd. | Non-invasive blood sugar measuring method and fingertip measuring probe |
| CN110338813A (en) * | 2019-06-04 | 2019-10-18 | 西安理工大学 | A kind of Noninvasive Blood Glucose Detection Methods based on spectrum analysis |
Non-Patent Citations (4)
| Title |
|---|
| ALEJANDRO等: "The fractional Fourier transform as a biomedical signal and image processing tool: A review", 《BIOCYBERNETICS AND BIOMEDICAL ENGINEERING》, 5 June 2020 (2020-06-05), pages 1081 - 1093 * |
| SHANTANU等: "In-Vivo Estimation of Glucose Level Using PPG Signal", 《2020 INTERNATIONAL CONFERENCE ON INFORMATION AND COMMUNICATION TECHNOLOGY CONVERGENCE （ICTC）》, 21 December 2020 (2020-12-21), pages 733 - 736 * |
| 凌振宝;侯洋;杨悦;赵鹏飞;: "便携式无创血糖检测方法", 吉林大学学报(信息科学版), no. 01, 15 January 2017 (2017-01-15), pages 63 - 67 * |
| 董金陇等: "脉搏信号伪基线去除方法研究", 《中国测试》, vol. 47, no. 4, 30 April 2021 (2021-04-30), pages 19 - 25 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Thatcher et al. | Quantitative EEG normative databases: Validation and clinical correlation | |
| CN109276241B (en) | Pressure identification method and equipment | |
| Yan et al. | The moment segmentation analysis of heart sound pattern | |
| US20180233227A1 (en) | Systems and methods of analyte measurement analysis | |
| Guo et al. | Analysis and recognition of traditional Chinese medicine pulse based on the hilbert-huang transform and random forest in patients with coronary heart disease | |
| KR20080082665A (en) | Systems and methods for analyzing and assessing depression and other mood disorders using electroencephalographic (eeg) measurements | |
| Ashok et al. | Determination of blood glucose concentration by using wavelet transform and neural networks | |
| CN112043252B (en) | Emotion recognition system and method based on respiratory component in pulse signal | |
| Stevenson et al. | A nonparametric feature for neonatal EEG seizure detection based on a representation of pseudo-periodicity | |
| CN111739647B (en) | Magnetic resonance image-based bipolar disorder suicide risk prediction method and related device | |
| Chen et al. | Short-term HRV in young adults for momentary assessment of acute mental stress | |
| US20220117552A1 (en) | Artificial neural network-based nuclear magnetic resonance and magnetic resonance spectroscopy data processing method and apparatus thereof | |
| CN112806977A (en) | Physiological parameter measuring method based on multi-scale fusion network | |
| KR102439319B1 (en) | Method and apparatus for processing nuclear magnetic resonance and magnetic resonance spectroscopy data based on artificial neural network | |
| CN117357080A (en) | Near infrared spectrum signal denoising method and device, terminal equipment and storage medium | |
| Jaros et al. | Novel phonocardiography system for heartbeat detection from various locations | |
| Friedman et al. | Angular offset distributions during fixation are, more often than not, multimodal | |
| CN116965812A (en) | Noninvasive blood glucose detection method and system based on fractional Fourier transform analysis | |
| Antunes et al. | A morphology-based feature set for automated Amyotrophic Lateral Sclerosis diagnosis on surface electromyography | |
| CN116451129A (en) | Pulse classification and identification method and system | |
| Geng et al. | Identification of major depression patients using machine learning models based on heart rate variability during sleep stages for pre-hospital screening | |
| US11324411B2 (en) | Method and system for determining cardiovascular parameters | |
| CN114557691A (en) | Non-invasive blood lipid detection method and system based on multi-wavelength PPG (photoplethysmography) signals | |
| TWI780396B (en) | Evaluation method and evaluation system of suicidal ideation based on multi-feature MRI and artificial intelligence | |
| CN113729708A (en) | Lie evaluation method based on eye movement technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |