CN115768348A - Device and equipment capable of realizing noninvasive blood glucose detection - Google Patents
Device and equipment capable of realizing noninvasive blood glucose detection Download PDFInfo
- Publication number
- CN115768348A CN115768348A CN202180024619.4A CN202180024619A CN115768348A CN 115768348 A CN115768348 A CN 115768348A CN 202180024619 A CN202180024619 A CN 202180024619A CN 115768348 A CN115768348 A CN 115768348A
- Authority
- CN
- China
- Prior art keywords
- principal component
- feature
- characteristic
- current
- blood glucose
- 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 424
- 210000004369 blood Anatomy 0.000 title claims abstract description 424
- 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 239
- 239000008103 glucose Substances 0.000 title claims abstract description 239
- 238000001514 detection method Methods 0.000 title claims abstract description 48
- 238000005259 measurement Methods 0.000 claims abstract description 99
- 230000007613 environmental effect Effects 0.000 claims abstract description 60
- 230000002503 metabolic effect Effects 0.000 claims abstract description 47
- 230000010354 integration Effects 0.000 claims abstract description 37
- 238000002329 infrared spectrum Methods 0.000 claims abstract description 34
- 239000013598 vector Substances 0.000 claims description 135
- 238000000034 method Methods 0.000 claims description 82
- 230000014509 gene expression Effects 0.000 claims description 79
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 62
- 238000012549 training Methods 0.000 claims description 58
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 40
- 229910052760 oxygen Inorganic materials 0.000 claims description 40
- 239000001301 oxygen Substances 0.000 claims description 40
- 238000012360 testing method Methods 0.000 claims description 29
- 238000010276 construction Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 17
- 230000036284 oxygen consumption Effects 0.000 claims description 15
- 238000013507 mapping Methods 0.000 claims description 13
- 208000015181 infectious disease Diseases 0.000 abstract description 5
- 238000005516 engineering process Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 30
- 230000004060 metabolic process Effects 0.000 description 16
- 230000006870 function Effects 0.000 description 15
- 239000011159 matrix material Substances 0.000 description 14
- 238000003860 storage Methods 0.000 description 14
- 238000000862 absorption spectrum Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 12
- 238000005457 optimization Methods 0.000 description 12
- 238000004590 computer program Methods 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 230000001419 dependent effect Effects 0.000 description 5
- 208000014674 injury Diseases 0.000 description 5
- 230000031700 light absorption Effects 0.000 description 5
- 238000010238 partial least squares regression Methods 0.000 description 5
- 238000006467 substitution reaction Methods 0.000 description 5
- 230000008733 trauma Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000004566 IR spectroscopy Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000017531 blood circulation Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 210000001367 artery Anatomy 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 206010012601 diabetes mellitus Diseases 0.000 description 2
- 230000008020 evaporation Effects 0.000 description 2
- 238000001704 evaporation Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000001766 physiological effect Effects 0.000 description 2
- 230000029058 respiratory gaseous exchange Effects 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 102000001554 Hemoglobins Human genes 0.000 description 1
- 108010054147 Hemoglobins Proteins 0.000 description 1
- 108010001267 Protein Subunits Proteins 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000036772 blood pressure Effects 0.000 description 1
- 230000036760 body temperature Effects 0.000 description 1
- 230000000747 cardiac effect Effects 0.000 description 1
- 230000004087 circulation Effects 0.000 description 1
- 235000005911 diet Nutrition 0.000 description 1
- 230000037213 diet Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 210000000624 ear auricle Anatomy 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035790 physiological processes and functions Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
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
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Optics & Photonics (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
An apparatus capable of non-invasive blood glucose detection comprises a parameter acquisition module (101), a feature determination module (102) and a blood glucose determination module (103). The parameter acquisition module (101) is used for acquiring physiological parameters of the blood sugar object to be tested and environmental parameters of the current region of the blood sugar object to be tested, and the physiological parameters are acquired in a non-invasive mode. The characteristic determination module (102) is used for obtaining a first input characteristic of the blood sugar object to be tested according to the physiological parameters and the environmental parameters, wherein the first input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the blood sugar object to be tested. The blood sugar determining module (103) is used for inputting the first input characteristic into the blood sugar measuring model so as to obtain the blood sugar value of the blood sugar object to be measured through the blood sugar measuring model. By adopting the device, the technical problems that the invasive blood sugar measurement mode in the related technology can not continuously monitor blood sugar and easily brings infection risk to patients can be solved.
Description
The embodiment of the application relates to the technical field of blood glucose detection, in particular to a device and equipment capable of realizing noninvasive blood glucose detection.
Blood glucose refers to the glucose in the blood. Measuring the blood glucose level in blood is of great importance for treating and observing diseases, for example, in the field of diabetes treatment, doctors can guide clinical medication of patients by monitoring the blood glucose level of patients and regulate daily diet movement of patients, thereby controlling diabetes. Invasive blood glucose measurements are typically used in monitoring blood glucose. In the invasive blood glucose measurement mode, blood of a patient needs to be collected first, and then the blood is analyzed to obtain the blood glucose value. Such a method does not allow continuous blood glucose monitoring and can cause pain and trauma to the patient during blood collection and can also risk infection.
Disclosure of Invention
The embodiment of the application provides a device and equipment capable of realizing noninvasive blood glucose detection, and aims to solve the technical problems that invasive blood glucose measurement modes in the related art cannot continuously monitor blood glucose and easily bring infection risks to patients.
In a first aspect, an embodiment of the present application provides an apparatus for enabling noninvasive blood glucose measurement, including: the device comprises a parameter acquisition module, a characteristic determination module and a blood sugar determination module;
the parameter acquisition module is used for acquiring physiological parameters of a blood sugar object to be tested and environmental parameters of an area where the blood sugar object to be tested is located currently, and the physiological parameters are acquired in a non-invasive mode;
the characteristic determination module is used for obtaining a first input characteristic of the blood sugar object to be tested according to the physiological parameter and the environmental parameter, wherein the first input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the blood sugar object to be tested;
the blood sugar determining module is used for inputting the first input characteristics into a blood sugar measuring model so as to obtain the blood sugar value of the object to be measured for blood sugar through the blood sugar measuring model.
In a second aspect, an embodiment of the present application further provides an apparatus capable of performing non-invasive blood glucose detection, including: the apparatus for enabling non-invasive blood glucose measurement as set forth in the first aspect.
The device and the equipment capable of realizing noninvasive blood glucose detection acquire physiological parameters of a blood glucose object to be detected acquired in a noninvasive mode and environmental parameters of a current area through the parameter acquisition module, then the characteristic determination module obtains first input characteristics corresponding to the blood glucose object to be detected according to the environmental parameters and the physiological parameters, then the blood glucose determination module inputs the first input characteristics to the blood glucose measurement model, and therefore the technical means for obtaining the blood glucose value of the blood glucose object to be detected through the blood glucose measurement model is achieved, and the technical problems that invasive blood glucose measurement modes in the related art cannot continuously monitor blood glucose and infection risks are easily brought to patients are solved. The physiological parameters of the blood sugar object to be detected are collected in a non-invasive mode, pain and trauma are not required to be brought to the blood sugar object to be detected, and continuous blood sugar value detection can be achieved. The infrared spectrum characteristic and the metabolic heat integration characteristic are used as input characteristics to be input into the blood glucose measurement model, the richness of the characteristics is improved, the two characteristics are mutually compensated, the robustness of the blood glucose measurement model is enhanced, the problems that the blood glucose measurement model is easily interfered by physiological tissues and environments when the infrared spectrum method or the metabolic heat integration method is used independently are solved, and the accuracy of the blood glucose value is ensured.
FIG. 1 is a schematic structural diagram of an apparatus for non-invasive blood glucose measurement according to an embodiment of the present application;
FIG. 2 is a flow chart of a method for performing non-invasive blood glucose monitoring according to an embodiment of the present application;
FIG. 3 is a flow chart of another method for enabling non-invasive blood glucose monitoring according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an apparatus for performing non-invasive blood glucose measurement according to an embodiment of the present application.
The present application will be described in further detail with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are for purposes of illustration and not limitation. It should be further noted that, for the convenience of description, only some of the structures related to the present application are shown in the drawings, not all of the structures.
The infrared spectroscopy and the metabolic heat integration method belong to two noninvasive blood glucose measurement methods. The principle of the infrared spectroscopy is that the concentration of glucose solution in blood has obvious correlation with the infrared absorption spectrum of a human body, and the blood glucose value can be analyzed through the infrared absorption spectrum, but physiological substances and tissues such as water, hemoglobin and fat in blood have absorption and scattering effects on light, so that the accuracy of the infrared absorption spectrum of glucose in blood can be influenced, and the accuracy of blood glucose measurement is further influenced. The principle of the metabolic heat integration method is that the heat generated by the metabolism of a patient has positive correlation with the concentration of glucose in blood, and the blood glucose value can be calculated through the metabolic heat, the blood flow rate and the blood oxygen saturation in the metabolic process of the patient. However, the metabolic heat integration method requires that the patient is in a state of static equilibrium, and if the patient is moving or the temperature and/or humidity of the environment changes, the amount of heat generated by metabolism is affected, and thus the accuracy of blood glucose measurement is affected.
Therefore, an embodiment of the present application provides an apparatus capable of performing non-invasive blood glucose detection, so as to improve accuracy of blood glucose level detection when performing non-invasive blood glucose detection by the apparatus.
The apparatus capable of performing non-invasive blood glucose detection provided in an embodiment of the present application may be implemented by a device capable of performing non-invasive blood glucose detection, where the device capable of performing non-invasive blood glucose detection may be formed by two or more physical entities, or may be formed by one physical entity, and the embodiment is not limited thereto.
Fig. 1 is a schematic structural diagram of an apparatus for performing non-invasive blood glucose measurement according to an embodiment of the present application. Referring to fig. 1, the apparatus capable of non-invasive blood glucose detection includes a parameter acquisition module 101, a feature determination module 102, and a blood glucose determination module 103.
The parameter acquisition module is used for acquiring physiological parameters of a blood sugar object to be detected and environmental parameters of an area where the blood sugar object to be detected is located currently, and the physiological parameters are acquired in a non-invasive mode; the characteristic determination module is used for obtaining a first input characteristic of the blood sugar object to be tested according to the physiological parameters and the environmental parameters, wherein the first input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the blood sugar object to be tested; and the blood sugar determining module is used for inputting the first input characteristic into the blood sugar measuring model so as to obtain the blood sugar value of the blood sugar object to be measured through the blood sugar measuring model.
In one embodiment, the blood glucose test object refers to a user who needs to measure blood glucose value, and the number of the blood glucose test objects may be one or more, which is not limited in the embodiment.
The current region of the blood glucose test object refers to a physical space where the blood glucose test object is currently located, and the physical space may be an indoor space or an outdoor space, which is not limited in the embodiment. The environmental parameter refers to a parameter related to the environment of the region where the blood glucose test object is located. In one embodiment, the environmental parameters include an ambient temperature and an ambient humidity, wherein the ambient temperature represents a degree of warmth and warmth of the area, and the ambient humidity represents a degree of dryness of the area. In one embodiment, the parameter obtaining module may communicate with the means for collecting the environmental parameter to obtain the environmental parameter. Wherein, the device for gathering the environmental parameter can be selected according to actual conditions to different environmental parameters can be gathered by different devices. In one embodiment, the device for non-invasive blood glucose monitoring is configured with a device for collecting environmental parameters, such as a temperature sensor and a humidity sensor, and collects the environmental temperature through the temperature sensor and the environmental humidity through the humidity sensor, and then the parameter acquiring module can acquire the collected environmental temperature and environmental humidity.
The physiological parameter refers to a parameter indicating a current physiological state of the subject to be measured, and in one embodiment, the physiological parameter includes a body surface temperature, a blood oxygen saturation level, an Electrocardiograph (ECG) signal, and a multi-channel pulse signal (PPG). The body surface temperature refers to the temperature of the skin surface in a blood sugar object to be measured, the blood oxygen saturation refers to the concentration of blood oxygen in blood, the electrocardiosignals are electric activity change graphs generated in each cardiac cycle of a heart, the pulse signals are change graphs representing the pulse of an artery, and the multichannel pulse signals refer to pulse signals acquired through a plurality of channels. In one embodiment, the physiological parameters are obtained in a non-invasive manner, i.e., the physiological parameters can be obtained without causing pain or trauma to the subject to be tested for blood glucose. For example, the physiological parameter may be acquired by the device for acquiring the physiological parameter and then sent to the parameter acquiring module, wherein the device for acquiring the physiological parameter may be selected according to actual conditions. In one embodiment, the apparatus for non-invasive blood glucose detection is configured with a device for acquiring physiological parameters, for example, the apparatus for non-invasive blood glucose detection is configured with an electrocardiograph acquisition device and an infrared acquisition device, wherein the electrocardiograph acquisition device acquires electrocardiograph signals by using electrocardiograph electrodes, i.e., the electrocardiograph electrodes are attached to a set position of a human body to acquire the electrocardiograph signals. The infrared acquisition device alternately irradiates an acquisition area (such as a fingertip or an earlobe) of the object to be measured for blood sugar by using a light source with a visible red light spectrum and a light source with an infrared spectrum, and further obtains body surface temperature, blood oxygen and multi-channel pulse signals for the absorption spectrum of the irradiation light source through the acquisition area. After the acquisition, the acquired physiological parameters are acquired by a parameter acquisition module. It should be noted that the above-mentioned infrared acquisition device adopts multiple channels, one channel is infrared light with a wavelength of 960nm, and the other channel is red light with a wavelength of 660nm, and the remaining channels are set to infrared light with other wavelengths, and different remaining channels can correspond to different wavelengths, so as to obtain corresponding physiological parameters according to the absorption spectra of multiple channels. The wavelength of the infrared light used in each of the remaining channels may be selected according to actual conditions, for example, the wavelength of the infrared light in one channel is 850nm, and the wavelength of the infrared light in the other channel is 760nm. It should be noted that, in practical applications, the physiological parameter may also include other contents, for example, the current physiological parameter includes a blood flow rate of the blood glucose test object. Optionally, in the embodiment, when the physiological parameter is collected, the blood glucose test object is in a static state or a non-strenuous exercise state, where the non-strenuous exercise state is a state that does not affect the physiological parameter.
In an exemplary embodiment, the parameter obtaining module obtains a physiological parameter and an environmental parameter, and then sends the currently obtained physiological parameter and the environmental parameter to the feature determining module, and the feature determining module is configured to analyze the environmental parameter and the physiological parameter to obtain a plurality of features related to the blood glucose value. It is understood that the first input feature is a feature relating to a blood glucose level, and the blood glucose level of the subject to be tested can be obtained by the first input feature. In one embodiment, the first input features include infrared spectroscopy features and metabolic heat conformation features. The infrared spectrum characteristic refers to the characteristic related to the infrared absorption spectrum, the principle of the infrared spectrum characteristic is that the concentration of the glucose solution has obvious correlation with the infrared absorption spectrum, and the blood glucose value can be calculated through the characteristic related to the infrared absorption spectrum. The infrared spectral signature may be derived from a physiological parameter (in one embodiment, a multichannel pulse signal). The metabolic heat integration characteristic is a characteristic related to the calorie generated by human metabolism obtained through environmental parameters and physiological parameters, and the blood sugar value can be obtained according to the principle that the blood sugar value has positive correlation with the calorie generated by human metabolism.
And after the characteristic determining module obtains the first input characteristic, the first input characteristic is sent to the blood sugar measuring module. The blood sugar measuring module is used for obtaining the blood sugar value of the blood sugar object to be measured according to the first input characteristic. The blood sugar measuring module can be realized through a blood sugar measuring model, namely, the blood sugar measuring model is deployed in the blood sugar measuring module so as to obtain the blood sugar value of the object to be measured for blood sugar through the blood sugar measuring model.
Wherein the blood sugar measurement model is a pre-constructed model for measuring blood sugar values. The construction basis of the blood sugar measurement model can be set according to the actual situation. In one embodiment, a blood glucose measurement model is constructed using a neural network, and then trained and deployed for application in a blood glucose determination module after training is complete. The specific structural embodiment of the neural network is not limited. The training process of the blood glucose measurement model comprises the steps of collecting infrared spectrum characteristics and metabolic heat integration characteristics corresponding to a plurality of blood glucose known objects, then taking the infrared spectrum characteristics and the metabolic heat integration characteristics as input, and training the blood glucose measurement model according to the blood glucose values of the blood glucose known objects so that the blood glucose values output by the blood glucose measurement model are close to the blood glucose values of the blood glucose known objects. In another embodiment, the blood glucose measurement model is constructed by a least square method, in which unknown data (blood glucose values obtained from the first input feature in the embodiment) can be easily obtained by the least square method, and the sum of squares of errors between these obtained data and actual data is minimized. At this time, the blood glucose measurement model may be constructed by collecting infrared spectrum characteristics and metabolic heat integration characteristics corresponding to the known blood glucose object, then calculating by using a least square method according to the infrared spectrum characteristics and the metabolic heat integration characteristics and blood glucose values corresponding to the known blood glucose object to obtain a regression equation, and using the regression equation as a regression equation of the blood glucose measurement model to complete construction of the blood glucose measurement model, wherein the regression equation may reflect a mathematical expression of a regression relationship of one variable (dependent variable) to another variable or a group of variables (independent variables), and the relationship between the blood glucose values and the infrared spectrum characteristics and the metabolic heat integration characteristics may be reflected by the regression equation. In yet another embodiment, the blood glucose measurement model is constructed using partial least squares regression. The partial least square regression method can also minimize the sum of squares of errors between the obtained data and the actual data, and in the partial least square regression method, principal components of the obtained data and the actual data (namely independent variables and dependent variables) are simultaneously obtained, so that the correlation between the principal components of the obtained data and the actual data is maximized, and then a regression equation between the principal components of the dependent variables and the principal components of the independent variables is obtained. And then, the regression equation is used as the regression equation of the blood glucose measurement model to complete the construction of the blood glucose measurement model.
After the blood sugar measurement model is constructed, the blood sugar measurement model is deployed in a blood sugar determination module. In application, after receiving the first input feature sent by the feature determination module, the blood glucose determination module inputs the first input feature into the blood glucose measurement model, and obtains a blood glucose value output by the blood glucose measurement model, where the blood glucose value is a blood glucose value of a blood glucose object to be measured. Optionally, the apparatus for non-invasive blood glucose detection further comprises a display device, and the display device can measure the obtained blood glucose value.
It can be understood that, when there are a plurality of blood glucose test objects, the first input features of the blood glucose test objects can be simultaneously input into the blood glucose measurement model, so as to simultaneously output blood glucose values of the blood glucose test objects through the blood glucose measurement model.
It can be understood that when the blood sugar value of the blood sugar test object needs to be continuously monitored, the physiological parameters and the environmental parameters can be continuously obtained by the device capable of realizing the non-invasive blood sugar detection in the above manner, and then the blood sugar value of the blood sugar test object can be continuously obtained.
In one embodiment, the environmental parameters include an environmental temperature and an environmental humidity, the physiological parameters include a body surface temperature, a blood oxygen saturation level, an electrocardiogram signal and a multichannel pulse signal, the metabolic heat conformation feature includes at least one of a temperature difference feature, a water vapor pressure difference feature, a blood oxygen saturation level feature, a pulse rate feature and a blood oxygen consumption feature per unit time, and the infrared spectrum feature includes a dominant frequency energy feature and/or a dominant frequency feature.
In one embodiment, the characteristic determination module comprises a metabolic heat integration characteristic determination unit and an infrared spectrum characteristic determination unit, wherein the metabolic heat integration characteristic determination unit is used for obtaining a temperature difference characteristic according to the body surface temperature and the environment temperature; and/or acquiring a first water vapor pressure corresponding to the body surface temperature and a second water vapor pressure corresponding to the environment temperature, and acquiring a water vapor pressure difference characteristic according to the first water vapor pressure, the second water vapor pressure and the environment humidity; and/or, using the blood oxygen saturation as a blood oxygen saturation characteristic; and/or obtaining pulse rate characteristics according to the multi-channel pulse signals; and/or obtaining the blood oxygen consumption characteristics in unit time according to the blood oxygen saturation, the multichannel pulse signals and the electrocardiosignals. The infrared spectrum characteristic determining unit is used for taking the main frequency energy of the multi-channel pulse signal as a main frequency energy characteristic; and/or, taking the main frequency of the multi-channel pulse signal as the main frequency characteristic.
The amount of heat generated by metabolism is related to factors such as temperature (ambient temperature and body surface temperature), humidity, blood oxygen saturation, blood oxygen consumption and pulse rate, and thus, in one embodiment, the set metabolic heat integration characteristics include at least one of a temperature difference characteristic, a water vapor pressure difference characteristic, a blood oxygen saturation characteristic, a pulse rate characteristic and a blood oxygen consumption characteristic per unit time. In one embodiment, the metabolic heat conformation features include a temperature difference feature, a water vapor pressure difference feature, a blood oxygen saturation feature, a pulse rate feature, and a blood oxygen consumption feature per unit time. The metabolic heat integration characteristic can be obtained by the metabolic heat integration characteristic determining unit, and it can be understood that the metabolic heat integration characteristic determining unit can obtain the corresponding physiological parameter and the environmental parameter obtained by the parameter obtaining module, and then obtain the metabolic heat integration characteristic according to the physiological parameter and the environmental parameter.
For example, when a human body feels cold, the body can resist the cold by accelerating the heat emitted by metabolism, and when the human body feels hot, the body can also accelerate the metabolism to consume much heat. Therefore, the influence of the temperature on the metabolism of the blood glucose test object can be determined by the temperature difference between the body surface temperature and the ambient temperature. In one embodiment, the temperature difference is recorded as a temperature difference signature. At this time, when the metabolic heat integration characteristic determining unit determines the temperature difference characteristic, the body surface temperature and the ambient temperature acquired by the parameter acquiring module are acquired, and the temperature difference characteristic is acquired according to the body surface temperature and the ambient temperature. Wherein the calculation mode of the temperature difference characteristic is f 1 ＝t s -t a Wherein f is 1 Characteristic of temperature difference, t s Indicating the body surface temperature, t a Representing the ambient temperature.
Illustratively, during the metabolic process, water vapor is emitted from the skin surface of the subject to be tested for blood glucose. The amount of evaporation from the skin surface of a subject with blood glucose is related to the pressure of water vapor in the air. For example, the greater the pressure of water vapor in the air, the less the amount of evaporation from the skin surface and the less heat that can be removed. The water vapor pressure is generally related to temperature and humidity, and in one embodiment, the effect of water vapor pressure on the metabolism of the blood glucose test object is determined by ambient humidity, ambient temperature, and body surface temperature. In one embodiment, the water vapor pressure on the skin surface of the blood glucose test object is recorded as a first water vapor pressure, and the water vapor pressure in the physical space is recorded as a second water vapor pressure. In one embodiment, the first water vapor pressure and the second water vapor pressure are both expressed by absolute pressures and the corresponding absolute pressures are different at different temperatures, such that the first water vapor pressure can be determined by a body surface temperature and the second water vapor pressure can be determined by an ambient temperature. In one embodiment, the temperature versus absolute pressure relationship is shown in Table 1 below.
| Temperature (degree centigrade) | Absolute pressure (kPa) | Temperature (centigrade) | Absolute pressure (kPa) |
| 20 | 2.3388 | 21 | 2.4877 |
| 22 | 2.6447 | 23 | 2.8104 |
| 24 | 2.985 | 25 | 3.169 |
| 26 | 3.3629 | 27 | 3.567 |
| 28 | 3.7818 | 29 | 4.0078 |
| 30 | 4.2455 | 31 | 4.4953 |
| 32 | 4.7578 | 33 | 5.0335 |
| 34 | 5.3229 | 35 | 5.6267 |
| 36 | 5.9453 | 37 | 6.2795 |
| 38 | 6.6298 | 39 | 6.9969 |
TABLE 1
Referring to table 1, absolute pressures corresponding to 20 ℃ to 39 ℃ are shown. At this time, the first water vapor pressure and the second water vapor pressure can be obtained from table 1. And then, obtaining the water vapor pressure difference by utilizing the first water vapor pressure, the second water vapor pressure and the environment humidity, and taking the water vapor pressure difference as the water vapor pressure difference characteristic. In one embodiment, the pressure difference characteristic of the water vapor is calculated by f 2 ＝p st -rh×p at Wherein f is 2 Characteristic of pressure difference of water vapor, p st Denotes the first water vapor pressure, p at Representing a second water vapor pressure and rh representing ambient humidity. It can be understood that the absolute pressure is multiplied by the corresponding humidity to obtain the relative pressure, and therefore, in one embodiment, the second water vapor pressure is multiplied by the ambient humidity to obtain the relative water vapor pressure of the region where the blood glucose test object is located. For example, the first water vapor pressure is determined as the relative water vapor pressure on the skin surface of the blood glucose test object, that is, since the humidity on the skin surface of the blood glucose test object is not obtained in the current embodiment, the humidity on the skin surface is defaulted to be 1, and the relative water vapor pressure of the blood glucose test object can be obtained by multiplying 1 by the first water vapor pressure. It can be understood that, after the device capable of realizing non-invasive blood glucose detection collects the humidity on the skin surface of the blood glucose test object, the first water vapor pressure in the above formula needs to be multiplied by the collected humidity first to obtain the relative water vapor pressure on the skin surface of the blood glucose test object, and at this time, the calculation formula of the water vapor pressure difference characteristic can be f 2 ＝rh’×p st -rh×p at Wherein rh' represents the humidity of the skin surface of the blood glucose test object. Then, after the relative water vapor pressure on the skin surface of the blood sugar object to be measured is subtracted from the relative water vapor pressure of the environment, the water vapor pressure difference characteristic can be obtained, namely the water vapor pressure difference characteristic is the difference of the relative water vapor pressure. In this case, when the metabolic heat integration characteristic determining unit determines the water vapor pressure difference characteristic, the metabolic heat integration characteristic determining unit stores the above table 1 and the formulaAfter the body surface temperature, the environment temperature and the environment humidity acquired by the parameter acquisition module are acquired, the first water vapor pressure difference and the second water vapor pressure difference are determined by using the table 1, and then the water vapor pressure difference characteristic is obtained according to a calculation formula.
Illustratively, the blood oxygen saturation can be understood as the blood oxygen concentration, which is an important physiological parameter reflecting the functions of respiration and circulation, and measures the capability index of human blood carrying oxygen, and the oxygen in blood needs to be consumed in the metabolic process, so the blood oxygen saturation is related to the metabolism 3 = o, wherein, f 3 Indicating the blood oxygen saturation characteristics and o indicates the blood oxygen saturation. When the metabolic heat integration characteristic determining unit determines the blood oxygen saturation characteristic, the blood oxygen saturation acquired by the parameter acquiring module is directly acquired and used as the blood oxygen saturation characteristic.
The blood oxygen consumption characteristic per unit time is exemplarily referred to as a consumption value of oxygen in blood per unit time, and the consumption value of oxygen per unit time is related to metabolism due to consumption of oxygen in blood during metabolism. Wherein, the blood oxygen consumption characteristic in unit time is determined by the multichannel pulse signal, the electrocardio signal and the blood oxygen saturation. In one embodiment, a peak point of the multi-channel pulse signal and a peak point of the electrocardiograph signal are obtained, then the two peak points are subtracted to obtain a pulse wave arrival time, and the blood oxygen saturation is divided by the pulse wave arrival time to obtain a blood oxygen consumption characteristic in unit time. Wherein, the calculation formula of the blood oxygen consumption characteristic in unit time is f 4 ＝o/p' at Wherein, f 4 Denotes blood oxygen consumption characteristics per unit time, o denotes blood oxygen saturation, p' at Representing the pulse wave arrival time. When the metabolic heat integration characteristic determining unit determines the blood oxygen consumption characteristic, the multichannel pulse signal, the electrocardiosignal and the blood oxygen saturation acquired by the parameter acquiring module are acquired, and then the blood oxygen consumption characteristic is acquired by using the formula.
Exemplary ofWhen the human body performs basic physiological activities (i.e. blood circulation, respiration and constant body temperature), the lowest heat consumed can be approximately obtained by calculating the pulse rate and the blood pressure, and therefore, the pulse rate of the human body is related to the heat generated by the basic metabolism (i.e. when the basic physiological activities are performed) of the human body. In one embodiment, the pulse rate is characterized as the pulse rate. Since the multi-channel pulse signal can reflect the variation of the artery fluctuation, the pulse rate characteristic can be obtained through the multi-channel pulse signal, and the pulse rate characteristic can be expressed as f 5 ＝p r Wherein f is 5 Indicating the pulse rate characteristic, p r Indicating the pulse rate. At this time, when the metabolic heat conformation characteristic determining unit determines the pulse rate characteristic, the multi-channel pulse signal acquired by the parameter acquiring module is specifically acquired, and then the pulse rate characteristic is acquired by using the multi-channel pulse signal.
It will be appreciated that f is obtained 1 To f 5 Then, the metabolic heat conformation feature determination unit will f 1 To f 5 As a metabolic heat conformation feature and sent to the blood glucose determination module.
Illustratively, the infrared light absorption spectrum of the human body can be obtained by multi-channel pulse signals, wherein one infrared light absorption spectrum can be obtained by each channel, and the dominant frequency and dominant frequency of the infrared light absorption spectrum are important parameters of the infrared light absorption spectrum, so that in one embodiment, the set infrared spectrum characteristics comprise dominant frequency energy characteristics and/or dominant frequency characteristics. In one embodiment, the infrared spectral features are described as including a primary frequency energy feature and a primary frequency feature. When multi-channel pulse signals are collected, the main frequency energy and the main frequency of each channel can be obtained. In one embodiment, the main frequency energy and the main frequency of a first channel in the multi-channel pulse signal are respectively used as the main frequency energy characteristic and the main frequency characteristic, for example, a channel corresponding to infrared light with a wavelength of 960nm is used as the first channel, and at this time, the main frequency energy characteristic and the main frequency characteristic are calculated through the infrared light absorption spectrum signal of the channel. At this time, the dominant frequency energy characteristic can be expressed as: f. of 6 ＝power DF Wherein f is 6 Representing main frequency energy characteristics, power DF Representing the dominant frequency energy of the first channel. The dominant frequency characteristic may be denoted as f 7 ＝freq DF Wherein f is 7 Indicating the dominant frequency characteristic, freq DF Representing the dominant frequency of the first channel. When the infrared spectrum characteristic determining unit acquires the main frequency energy characteristic and the main frequency characteristic, the multi-channel pulse signals acquired by the parameter acquisition device are acquired, and then the main frequency energy characteristic and the main frequency characteristic are acquired according to the multi-channel pulse signals. In practical application, the main frequency energy and the main frequency of more channels can be obtained to obtain a plurality of main frequency energy characteristics and a plurality of main frequency characteristics.
Understandably, the above f is obtained 6 To f 7 Then, the infrared spectrum characteristic determination unit will f 6 To f 7 As infrared spectrum characteristics and input to the blood glucose determination module.
And then, the blood sugar determining module obtains the blood sugar value according to the infrared spectrum characteristic and the metabolic heat integration characteristic.
The physiological parameters of the blood sugar object to be measured and the environmental parameters of the current region acquired in a noninvasive mode are acquired through the parameter acquisition module, then the first input characteristics corresponding to the blood sugar object to be measured are acquired by the characteristic determination module according to the environmental parameters and the physiological parameters, and the first input characteristics are input into the blood sugar measurement model by the blood sugar determination module, so that the blood sugar value of the blood sugar object to be measured is acquired through the blood sugar measurement model. The physiological parameters of the blood sugar test object are collected in a non-invasive mode, pain and trauma are not needed to be brought to the blood sugar test object, and continuous blood sugar value detection can be achieved. The infrared spectrum characteristic and the metabolic heat integration characteristic are used as input characteristics to be input into the blood glucose measurement model, the richness of the characteristics is improved, the two characteristics are mutually compensated, the robustness of the blood glucose measurement model is enhanced, the problems that the blood glucose measurement model is easily interfered by physiological tissues and environments when the infrared spectrum method or the metabolic heat integration method is used independently are solved, and the accuracy of the blood glucose value is ensured.
It can be understood that the performance of the blood glucose measurement model can directly influence the blood glucose measurement result, and besides the application of the blood glucose measurement model, the construction process of the blood glucose measurement model is also an important link. In one embodiment, the apparatus capable of non-invasive blood glucose detection may further have a function of constructing a blood glucose measurement model. It can be understood that, in practical application, the blood glucose measurement model may be further constructed by other apparatuses, and after the construction is completed, the blood glucose measurement model is deployed by other apparatuses in the blood glucose determination module of the apparatus capable of realizing the non-invasive blood glucose detection, so as to implement the application of the blood glucose measurement model. In one embodiment, the apparatus for non-invasive blood glucose measurement is described as an example of constructing a blood glucose measurement model. At this time, the device capable of realizing non-invasive blood glucose detection further comprises a feature acquisition module and a model construction module.
The characteristic acquisition module is used for acquiring training characteristics and target characteristics, the training characteristics comprise second input characteristics of at least one object with known blood sugar, the second input characteristics comprise infrared spectrum characteristics and metabolic heat integration characteristics of the object with known blood sugar, and the target characteristics comprise blood sugar values of the at least one object with known blood sugar; and the model building module is used for building a blood sugar measurement model through the training characteristics and the target characteristics.
Illustratively, a subject with known blood glucose refers to a subject for whom a blood glucose value has been specified, and in one embodiment, the blood glucose value is the blood glucose concentration in the fingertip blood or venous blood of the subject with known blood glucose. When the blood sugar measurement model is constructed, the characteristics required by training are firstly obtained by the characteristic obtaining module, and then the blood sugar measurement module is constructed by the model construction module according to the characteristics required by training.
In one embodiment, after acquiring the physiological parameter and the environmental parameter of the subject with known blood sugar, the second input characteristic of the subject with known blood sugar can be obtained according to the physiological parameter and the environmental parameter. The number of the collected blood sugar known objects can be set according to actual conditions, and in one embodiment, physiological parameters and environmental parameters of a plurality of blood sugar known objects are collected for example. The blood sugar known objects can be located in the same area or different areas, and it can be understood that the area where the blood sugar known object is located can be understood as the physical space where the blood sugar known object is located.
In one embodiment, the physiological parameters of each blood glucose known object comprise body surface temperature, blood oxygen saturation, electrocardio signals and multichannel pulse signals, and the environmental parameters comprise environmental temperature and environmental humidity. At this time, the feature acquisition module may integrate a device for acquiring the physiological parameter and the environmental parameter, or have the same function as the device for acquiring the physiological parameter and the environmental parameter, so as to acquire the physiological parameter and the environmental parameter of the blood glucose known object, and the feature acquisition module may further integrate a parameter acquisition module and a feature determination module, or have the same function as the parameter acquisition module and the feature determination module, so as to acquire the physiological parameter and the environmental parameter and obtain the corresponding second input feature. Or, the feature obtaining module may communicate with a device for collecting the physiological parameter and the environmental parameter to obtain the physiological parameter and the environmental parameter, and then obtain the corresponding second input feature.
In one embodiment, the description is given by taking an example that the feature obtaining module obtains the second input feature according to the physiological parameter and the environmental parameter. In an embodiment, the features corresponding to each blood glucose known object are recorded as second input features, the second input features are features used in constructing a blood glucose measurement model, the second input features and the first input features include features of the same kind, that is, the second input features also include infrared spectrum features and metabolic heat integration features, and the determination manner of the second input features can refer to the determination manner of the first input features (that is, the feature acquisition module can be integrated with the feature determination module to acquire the second input features according to the environmental parameters and the physiological parameters through the feature determination module, or the feature acquisition module has the same functions as the feature determination module to determine the second input features according to the working manner of the feature determination module). And then integrating the second input features into training features by the feature acquisition module. In one embodiment, the training features are in the form of a matrix, and each second input feature is a row vector in the matrix. In practical application, it can be understood that other devices independent of the device capable of non-invasive blood glucose detection may be used to collect the physiological parameters and the environmental parameters and determine the second input features, and then the second input features are sent to the feature acquisition module, and the feature acquisition module integrates the second input features into the training features.
The target characteristics consist of blood sugar values of the blood sugar known objects, and the characteristic acquisition module can directly acquire the input blood sugar values of the blood sugar known objects. Illustratively, the feature obtaining module obtains a second input feature and a blood glucose value of each blood glucose known object in sequence, obtains a training feature according to the second input feature, and obtains a target feature according to the blood glucose value. In one embodiment, the target feature is in the form of a vector, with one element in the vector representing the blood glucose value of a subject with known blood glucose. And the blood sugar value described by the second element in the target characteristic corresponds to the second input characteristic in the second row of the training characteristic corresponding matrix, and the analogy is repeated to ensure that the second input characteristic in the training characteristic corresponds to the blood sugar value in the target characteristic.
For example, the feature obtaining module obtains n second input features of the subjects with known blood glucose and blood glucose values, where the second input features of each of the subjects with known blood glucose include k features, and at this time, the training features are n × k matrices, and the target features are n × 1 vectors.
Illustratively, the model building module builds the blood glucose measuring module in a partial least squares regression manner, wherein the partial least squares regression manner is specifically used to determine a relational expression between the training features and the target features when the blood glucose measuring module is built in the partial least squares regression manner, so as to determine how to obtain corresponding blood glucose values through the second input features, and further build the blood glucose measuring model through the relational expression.
In one embodiment, when the model construction module constructs the blood glucose measurement model, a partial least square regression method is used for determining a regression equation between the training features and the target features when the correlation between the principal components of the training features and the target features is maximum and the variance of each principal component is maximum, so that the regression equation is used as an equation used by the blood glucose measurement model, and the blood glucose measurement model is constructed. Wherein a group of variables (such as training features or target features in the embodiment) with possible correlation is converted into a group of linearly uncorrelated variables by orthogonal transformation, and the converted group of variables can be understood as principal components. In one embodiment, when the model construction module constructs the blood glucose measurement model by using partial least square regression, the model construction module comprises a current feature determination unit, a first principal component axis vector determination unit, a first residual error determination unit, a second residual error determination unit, an updating unit, a linear expression construction unit and a model construction unit, wherein the current feature determination unit is used for taking a training feature as a current first feature and taking a target feature as a current second feature; the first principal component axis vector determining unit is used for obtaining a first principal component axis vector of the current first characteristic according to the current first characteristic and the current second characteristic; a first residual determining unit, configured to determine, according to the first principal component axis vector, a first principal component of the current first feature and a first residual required when the first principal component and the first principal component axis vector represent the current first feature; a second residual determining unit configured to determine, according to the first principal component and the current second feature, a target coefficient and a second residual that are required when the current second feature is represented by the first principal component; the updating unit is used for updating the first residual error into a current first characteristic, updating the second residual error into a current second characteristic, and returning to execute the operation of obtaining a first principal component axis vector of the current first characteristic according to the current first characteristic and the current second characteristic until an iteration stop condition is met; the linear expression construction unit is used for constructing a linear expression of the target feature relative to the training feature by using the first principal component axis vector, the first principal component and the target coefficient obtained by each iteration; and the model building unit is used for taking the linear expression as a regression equation used by the blood glucose measurement model so as to complete the building of the blood glucose measurement model.
In order to facilitate the description of the process of constructing the blood glucose measurement model, in one embodiment, the training features are first used as current first features, and the target features are used as current second features, where the current first features are in a matrix form and the second features are in a vector form. And the subsequent processing procedures are described by the current first characteristic and the current second characteristic. This process may be implemented by the current feature determination unit.
For example, after the current first feature and the current second feature are respectively mapped to the corresponding principal component axis vectors, the corresponding principal components can be obtained. The principal component axis vector can be understood as a coefficient used for determining the principal component. In one embodiment, a principal component axis vector corresponding to a current first feature is taken as a first principal component axis vector, a principal component axis vector corresponding to a current second feature is taken as a second principal component axis vector, a principal component corresponding to a current first feature is taken as a first principal component, a principal component corresponding to a current second feature is taken as a second principal component, and the first principal component and the second principal component are both expressed by vectors. The first principal component is obtained by multiplying the current first characteristic by the first principal component axial vector, and the second principal component is obtained by multiplying the current second characteristic by the second principal component axial vector.
In one embodiment, the first principal component axis vector determination unit takes the maximum correlation between the first principal component and the second principal component, the maximum variance of the first principal component, and the maximum variance of the second principal component as optimization targets, and solves the first principal component axis vector and the second principal component axis vector so that the first principal component and the second principal component can satisfy the optimization targets when the first principal component and the second principal component are obtained by the first principal component axis vector and the second principal component axis vector. In one embodiment, the first principal component axis vector and the second principal component axis vector are solved by a lagrange multiplier method. The lagrange multiplier method is also understood to be a lagrange multiplier method. Illustratively, the first principal component axis vector determination unit includes a mapping subunit and an axis vector determination subunit.
The mapping subunit is configured to map the current first feature to the first candidate principal component axis vector to obtain a first candidate principal component of the current first feature, and map the current second feature to the second candidate principal component axis vector to obtain a second candidate principal component of the current second feature; and the axis vector determining subunit is configured to determine, by using a lagrangian multiplier method, a first candidate principal component axis vector and a second candidate principal component axis vector when the first candidate principal component and the second candidate principal component are optimal, and use the first candidate principal component axis vector when the first candidate principal component is optimal as a first principal component axis vector of the current first feature.
For example, the first candidate principal component axis vector may be understood as an initial first principal component axis vector pre-established in the mapping subunit, and may also be understood as a first principal component axis vector to be solved. The second candidate principal component axis vector may be understood as an initial second principal component axis vector pre-established in the mapping subunit, and may also be understood as a second principal component axis vector to be solved. It can be understood that the first candidate principal component axis vector and the second candidate principal component axis vector can be optimized subsequently by using a lagrange multiplier method to obtain a final first principal component axis vector and a final second principal component axis vector.
The first candidate principal component is a principal component obtained by the mapping subunit through the first candidate principal component axis vector, and the second candidate principal component is a principal component obtained by the mapping subunit through the second candidate principal component axis vector. With the optimization of the first candidate principal component axis vector and the second candidate principal component axis vector, the correlation between the first candidate principal component and the second candidate principal component becomes maximized and the respective variances become maximized to achieve the optimization goal.
Illustratively, the first candidate principal component and the second candidate principal component are optimal, which means that the optimization goal is reached, i.e., the correlation between the first candidate principal component and the second candidate principal component is maximized and the respective variance is maximized.
In one embodiment, the first principal component axis vector and the second principal component axis vector are solved by using a lagrange multiplier method, and the specific implementation process is as follows:
by X representing the current first feature, y representing the current second feature, w 1 Representing the first candidate principal component axis vector, w 1 Is a k × 1 vector, k is the number of features contained in the line vector in the current first feature, c 1 Representing a second candidate principal component axis vector, c 1 Is a 1 × 1 vector, t 1 Denotes the first alternative principal component, u 1 Representing a second alternative principal component. Wherein, t 1 The calculation of (c) is referred to the following equation 1:
t 1 ＝X*w 1 (1)
u 1 the calculation of (c) is referred to the following equation 2:
u 1 ＝y*c 1 (2)
in one embodiment, the optimization objective is t 1 And u 1 Is maximized and t 1 And u 1 The respective variance is maximum, and in this case, the optimization objective can be expressed as: max < X w 1 ,y*c 1 ＞,s.t.||w 1 ||＝1,||c 1 I | =1. Wherein, | | w 1 I represents w 1 The length of the die, | | c 1 I means c 1 Die length of (2).
Then, when performing the lagrangian multiplier method, introducing a lagrangian multiplier λ and a lagrangian multiplier θ, and at this time, converting the optimization target into the following formula 3:
wherein the variables of equation 3 include w 1 And c 1 ，w 1 T Is w 1 Transpose of (2), X T Is the transpose of X, c 1 T Is c 1 The transposing of (1).
Then, for w in equation 3 1 Taking the derivative and making the derivative result 0, equation 4 is obtained:
at the same time, for c in formula 3 1 Taking the derivative and making the derivative result 0, equation 5 is obtained:
wherein, y T Is the transpose of y.
Then, the variable c is multiplied to each of the two sides of the right equal sign in equation 4 1 Multiplying the two sides of the left equal sign in equation 5 by the variable w 1 Then, according to the optimization target | | | w 1 ||＝1,||c 1 The equation 6 can be obtained by | = 1:
in this case, as can be seen from equation 6, the lagrange multiplier λ = θ.
After that, substituting equation 4 into equation 5, and introducing λ = θ during the substitution, equation 7 can be obtained:
X T yy T Xw 1 ＝λ 2 w 1 (7)
meanwhile, substituting formula 5 into formula 4, and introducing λ = θ during the substitution process may result in formula 8:
y T X T Xyc 1 ＝λ 2 c 1 (8)
then, according to formula 7 and formula 8, the correlation of the first candidate principal component and the second candidate principal component in the optimization objective is less than X × w 1 ,y*c 1 Is converted to Lambda max, < X w 1 ,y*c 1 Can be converted to equation 9:
＜X*w 1 ,y*c 1 ＞＝w 1 T X T yc 1 ＝w 1 T (λλ 1 )＝λ (9)
from equation 9, in conjunction with equation 7, it can be determined: w is a 1 Is a symmetric matrix X T yy T The unit feature vector corresponding to the maximum feature value of X, in conjunction with equation 8, may be determined: c. C 1 Is a symmetric matrix y T X T The unit feature vector corresponding to the maximum feature value of Xy. At this time, the axial vector determination subunit passes through the respective pairs X T yy T X and y T X T Xy is diagonalized to obtain a first candidate principal component axial vector w of X 1 And a second candidate principal component axis vector c of y 1 . Then, the calculated w 1 C is calculated as the first principal component axis vector of X 1 The second principal component axis vector as y.
For example, after the first principal component axis vector is obtained by the first principal component axis vector determining unit, the first principal component may be obtained by mapping the current first feature to the first principal component axis vector by the first residual error determining unit. It can be understood that there may be an error when the first principal component axis vector and the second principal component axis vector are determined by the lagrange multiplier method, and therefore, a residual error is introduced when the first current feature is represented by the first principal component and the first principal component axis vector, so as to accurately represent the current first feature. It can be understood that, since the current first feature is in a matrix form, the first residuals are also in a matrix form, and the matrix size of the first residuals is the same as the matrix size of the current first feature matrix.
In one embodiment, the first residual determining unit comprises a first principal component determining sub-unit, a first residual calculating sub-unit. The first principal component determining subunit is used for mapping the current first feature to the first principal component axis vector to obtain a first principal component of the current first feature; and the first residual error calculation subunit is used for determining a first residual error required when the first principal component and the first principal component axis vector represent the current first feature according to the first principal component and the first principal component axis vector.
It can be understood that the mapping of the current first feature to the first principal component axis vector by the first principal component determining subunit can refer to equation 1 to obtain the first principal component t 1 . The first principal component determining subunit maps the current second feature to the second principal component axis vector, referring to equation 2, to obtain a second principal component u 1 。
For example, when the current first feature is represented by the first principal component and the first principal component axis vector, the current first feature may be represented by equation 10:
X＝t 1 w 1 T +E (10)
where E is the first residual, it will be appreciated that in order to avoid w 1 There is an error such that t 1 w 1 T Not equal to X, so E is introduced. Wherein, w 1 、t 1 And X are both known, so the first residual calculation subunit can be given E by equation 10.
Optionally, the first residual calculation subunit may further use the second principal component and the second principal component axis vector to represent the current second feature, and in this case, the current second feature may be represented by equation 11:
y＝u 1 c 1 T +g (11)
wherein, g is a third residual, and the third residual has the same action as the first residual, which is not described herein again.
And then, the second residual error determining unit acquires the first principal component obtained by the first residual error determining unit, and determines a target coefficient and a second residual error according to the first principal component and the current second characteristic. It can be understood that the current second feature can be obtained by regression from the current first feature, that is, the current first feature is used as an independent variable, and the current second feature is used as a dependent variable, and the dependent variable can be changed by the independent variable. Meanwhile, the current first feature may be regressed from its own first principal component. Thus, the current second feature may be obtained by regression through the first principal component, i.e. the current second feature may be represented by the first principal component. At this time, the current second characteristic can be expressed as equation 12:
y＝t 1 r 1 T +f (12)
wherein r is 1 To use the first principal component to represent the coefficients used in the current second feature, in one embodiment, the coefficients are denoted as target coefficients, which are in the form of vectors. f is a residual used when the first principal component is used to represent the current second feature, and in one embodiment, the residual is referred to as a second residual, where the second residual is in the form of a vector. The second residual error determining unit passes t according to the above formula 1 And y calculating a target coefficient and a second residual. Wherein, the calculation mode of the target coefficient and the second residual error can be combined with actual selection.
In one embodiment, the target coefficient is calculated by a least squares method and the second residual is calculated by the target coefficient. At this time, the second residual determining unit includes a regression expression equation constructing subunit, a target coefficient determining subunit, and a second residual calculating subunit.
The regression expression equation constructing subunit is used for constructing a regression expression equation of the current second characteristic by using the first principal component, the target coefficient and the second residual error; a target coefficient determination subunit, configured to determine a target coefficient in the regression expression equation by a least square method; and the second residual error calculation subunit is used for substituting the first principal component and the target coefficient into the regression expression equation to obtain a second residual error.
The regression expression equation is an expression formula of the current second characteristic when the current second characteristic is obtained through regression of the first principal component, and the formula 12 is the regression expression equation of the current second characteristic and is obtained by constructing a subunit from the regression expression equation.
In one embodiment, when calculating the coefficient vector (i.e. the target coefficient) by the least square method, the coefficient vector can be calculated by equation 13, where equation 13 is:
wherein. I t 1 | is t 1 Die length of (2). Due to t 1 And y are known numbers, so the target coefficient determination subunit can calculate r by equation 13 1 。
Then, the second residual error calculation subunit determines the target coefficient r obtained by the subunit from the target coefficient 1 The first principal component t obtained by the first residual error determination unit 1 And substituting the current second characteristic y obtained by the current characteristic determining unit into a regression expression equation (namely formula 12) obtained by the regression expression equation constructing subunit to obtain a second residual error f.
After that, the updating unit judges whether or not the iteration stop condition is satisfied. If the iteration stop condition is not met, the updating unit updates the first residual error to be the current first feature, updates the second residual error to be the current second feature, and inputs the current first feature and the current second feature to the first principal component axis vector determining unit for calculation again. If the iteration stopping condition is met, the updating unit sends the first principal component axis vector, the first principal component and the target coefficient obtained by each iteration to the linear expression building unit, so that the linear expression building unit builds a linear expression of the target feature relative to the training feature.
Illustratively, the operation process of the update unit is as follows: after the first residual error and the second residual error are obtained, it is determined that the iteration process is finished, and it can be understood that the first residual error and the second residual error can be obtained in each iteration process. The iteration has the effect that the first residual error and the second residual error obtained by each calculation gradually become smaller, and finally the first residual error and the second residual error gradually tend to 0. In one embodiment, after the iteration process is finished, whether the iteration stop condition is currently met is judged. The iteration stop condition is a condition for stopping the iteration process, and the specific content thereof may be set according to actual conditions. In one embodiment, the iteration stop condition includes that the current iteration number reaches a target iteration number, and a specific value of the target iteration number may be set according to an actual situation. For example, the target number of iterations is 3, and as another example, the target number of iterations is equal to or greater than 1 and equal to or less than the total number of feature types included in the infrared spectrum feature and the metabolic heat conformation feature. After the first residual error and the second residual error are obtained each time, updating the current iteration times, wherein the current iteration times refer to the times of current iteration, then judging whether the current iteration times reach the target iteration times, and if the current iteration times reach the target iteration times, determining that the current first residual error and the current second residual error are small enough and do not need to be reduced again, so that the condition of stopping the iteration is determined to be met. Otherwise, determining that the iteration stop condition is not met, updating the first residual error into the current first characteristic, updating the second residual error into the current second characteristic, and inputting the current first characteristic and the current second characteristic into the first principal component axis vector determination unit for calculation again.
For example, the above process of reducing the first residual and the second residual may be regarded as a residual optimization process. It can be understood that, during optimization, the currently obtained first residual error may be updated to the current first feature, the currently obtained second residual error may be updated to the current second feature, and a new iteration process is started, that is, the current first feature and the current second feature are input to the first principal component axis vector determination unit to be calculated again. It can be understood that, after the first residual is taken as the current first feature, when iterative computation is performed again, the first principal component axis vector of the first residual may be obtained by the first principal component axis vector determining unit, the first principal component of the first residual may be obtained by the first residual determining unit, and a new first residual may be obtained when the first residual is represented by the first principal component axis vector and the first principal component. Similarly, the second residual error is used as the current second characteristic, and the target coefficient and the new second residual error obtained by the second residual error determining unit can also be obtained.
It can be understood that in the practical application of the apparatus capable of non-invasive blood glucose detection, the current feature determination unit and the updating unit are not required to be arranged, and only the iterative computation is required to be performed by the first principal component axis vector determination unit, the first residual error determination unit and the second residual error determination unit until the iteration stop condition is satisfied.
Illustratively, after the iteration is completed, the linear expression building unit obtains the first principal component axis vector and the first principal component obtained by each calculation. At this time, the linear expression building unit represents the training feature by the first principal component axis vector and the first principal component obtained at each iteration. For example, the training feature at the first iteration is taken as the current first feature, and the training feature after the iteration is completed may be represented as X = t 1 w 1 T +E 1 Where X represents a training feature, t 1 Representing a first principal component, w, obtained in a first iteration 1 Representing a first principal component axis vector, E, from a first iteration 1 Representing a first residual resulting from a first iteration. In the second iteration, the first residual error E is 1 As a current first feature, E is obtained after the iteration is completed 1 ＝t 2 w 2 T +E 2 Wherein, t 2 Representing the first principal component, w, obtained in the second iteration 2 Representing the first principal component axis vector, E, from the second iteration 2 Representing the first residual resulting from the second iteration. At this time, E 1 ＝t 2 w 2 T +E 2 Substitution X = t 1 w 1 T +E 1 After, the training feature may be represented as X = t 1 w 1 T +t 2 w 2 T +E 2 And the analogy is repeated until the training features are represented by the first principal component axis vector and the first principal component obtained by each iteration and the first residual error obtained by the last iteration after the iteration is finished. Similarly, the linear expression construction unit represents the target feature by the first principal component and the target coefficient obtained by each iteration. For example, in the first iteration, the target feature is used as the current second feature, and the target feature after the iteration is completed may be represented as y = t 1 r 1 T +f 1 Wherein y represents a target feature, t 1 Representing a first principal component, r, obtained in a first iteration 1 Representing the target coefficient, f, obtained in the first iteration 1 Representing a second residual resulting from the first iteration. In the second iteration, the second residual error f is 1 As a current second feature, f is obtained after the iteration is completed 1 ＝t 2 r 2 T +f 2 Wherein, t 2 Representing the first principal component, r, obtained in the second iteration 2 Representing the target coefficient, f, obtained in the second iteration 2 Representing a second residual resulting from the second iteration. At this time, f is 1 ＝t 2 r 2 T +f 2 Substitution y = t 1 r 1 T +f 1 After that, the target characteristic may be represented as y = t 1 r 1 T +t 2 r 2 T +f 2 And repeating the steps until the target characteristic is represented by the target coefficient and the first principal component obtained by each iteration and the second residual error obtained by the last iteration after the iteration is finished.
Because the training features and the target features can be both expressed by the first principal component, the linear expression building unit can substitute the training features into the expression of the target features through the relationship between the first principal component and the training features to obtain the expression of the target features relative to the training features, wherein the expression is a linear expression, that is, the target features and the training features satisfy a linear relationship.
As can be seen from the above, the linear expression building unit includes a first expression building subunit and a linear expression determining subunit. The first expression constructing subunit is used for constructing a first expression of the target characteristic by using the first principal component and the target coefficient obtained by each iteration; and the linear expression determining subunit is used for replacing the first principal component in the first expression by using the training feature and the first principal component axis vector obtained by each iteration to obtain a linear expression of the target feature relative to the training feature.
In one embodiment, the target iteration number is set to be n times, where n is greater than or equal to 2, and at this time, the linear expression of the target feature constructed by the first expression construction subunit is: y = t 1 r 1 T +t 2 r 2 T +···+t n r n T + f. Wherein y represents a target feature, t 1 A first principal component, t, obtained for the first iteration 2 The first principal component obtained for the second iteration process, and so on, t n For the first principal component, r, obtained in the nth iteration 1 For the target coefficient, r, obtained for the first iteration 2 For the target coefficient obtained in the second iteration process, and so on, r n And f is a second residual error obtained in the nth iteration process. Where f tends to 0.
Similarly, the first expression building subunit may also build a linear expression of the training features, where the linear expression is: x = t 1 w 1 T +t 2 w 2 T +···+t n w n T + E. Wherein X represents a training feature, t 1 For the first principal component, t, obtained in the first iteration 2 The first principal component obtained for the second iteration process, and so on, t n First principal component, w, obtained for the nth iteration 1 A first principal component axis vector, w, obtained for a first iteration process 2 The first principal component axis vector obtained for the second iteration process, and so on, w n And E is a first principal component axis vector obtained in the nth iteration process, and E is a first residual error obtained in the nth iteration process. Where E goes to 0.
In order to distinguish the linear expressions of the training features and the target features, in one embodiment, the linear expression of the target features is recorded as a first expression, and the linear expression of the training features is recorded as a second expression.
It can be understood that, as can be seen from the above expression, the blood glucose measurement model contains n principal components (t) 1 To t n ) In this case, the principal component matrix of the blood glucose measurement model is T = (T) 1 ,t 2 ···t n ) T . The blood glucose measurement model comprises n target coefficients (r) 1 To r n ) At this time, the target coefficient matrix of the blood glucose measurement model is R = (R) 1 ,r 2 ···r n ) T At this time, the first expression may be expressed as y = TR T + f. Meanwhile, the blood glucose measurement model also comprises n first principal component axial vectors (w) 1 To w n ) At this time, the principal component axis matrix of the blood glucose measurement model is W = (W) 1 ,w 2 ,···,w n ) T . Accordingly, the second expression may be expressed as X = TW T +E。
The process of the linear expression determination subunit determining the linear expression is as follows: exemplarily, X = TW T In + E, E approaches to 0, so T may be represented by X and W after ignoring E in one embodiment, i.e., T = XW, and then the first expression may be represented after substituting T = XW into the first expression: y = TR T +f＝XWR T +f＝Xa T + f wherein, a T ＝WR T . At this time, y = Xa T + f is the linear expression of the resulting target feature with respect to the training features.
Then, the model construction unit takes the linear expression constructed by the linear expression construction unit as a regression equation used by the blood glucose measurement model, namely the model construction unit deploys the regression equation of the blood glucose measurement model in the blood glucose determination module. Wherein, the regression equation of the blood sugar measurement model is as follows: y = Xa T + f at this time, a T Which may be understood as a regression coefficient vector used by the blood glucose measurement model.
In the subsequent application process, the characteristic determination module can calculate the blood sugar value y by substituting the first input characteristic into the X in the regression equation of the blood sugar determination module after acquiring the first input characteristic.
The device for realizing non-invasive blood glucose detection calculates the regression equation of the blood glucose measurement model through partial least square regression, further realizes the technical means for constructing the blood glucose measurement model, can realize that linear modeling (namely a linear expression of target characteristics about training characteristics) is carried out when the number of data samples (namely the number of physiological parameters and environmental parameters) is less than the number of input characteristics (namely the number of characteristics contained in infrared spectrum characteristics and metabolic heat integration characteristics), can greatly expand the number of the input characteristics, enrich the information content contained in the input characteristics, and enhance the accuracy of modeling. And when the blood glucose measurement model is constructed by using partial least square regression, the blood glucose measurement model can be expanded into a nonlinear model, namely, the relation between the training characteristics and the target characteristics is modified into a nonlinear relation, so that the blood glucose measurement model is expanded and modeled. Meanwhile, the characteristics required by the blood sugar measurement model are acquired in a non-invasive mode, pain and trauma are not required to be brought to a blood sugar object to be measured, the risk of infection is avoided, and continuous blood sugar value measurement can be realized.
It should be noted that, in the embodiment of the apparatus for non-invasive blood glucose detection, the units and modules included in the apparatus are merely divided according to the functional logic, but are not limited to the above division as long as the corresponding functions can be realized; in addition, specific names of the functional units are only used for distinguishing one functional unit from another, and are not used for limiting the protection scope of the present application.
Fig. 2 is a flowchart of a method for enabling noninvasive blood glucose measurement according to an embodiment of the present application, where the method for enabling noninvasive blood glucose measurement can be performed by the apparatus for enabling noninvasive blood glucose measurement. The method capable of realizing the non-invasive blood sugar detection is an application process of a blood sugar measurement model. Referring to fig. 2, the method for enabling noninvasive blood glucose measurement includes:
and step 210, acquiring physiological parameters of the blood sugar object to be detected and environmental parameters of the area where the blood sugar object to be detected is located currently, wherein the physiological parameters are acquired in a non-invasive mode.
And step 220, obtaining a first input characteristic of the blood sugar object to be tested according to the physiological parameter and the environmental parameter, wherein the first input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the blood sugar object to be tested.
In one embodiment, the environmental parameters include an environmental temperature and an environmental humidity, the physiological parameters include a body surface temperature, a blood oxygen saturation level, an electrocardiogram signal and a multichannel pulse signal, the metabolic heat conformation feature includes at least one of a temperature difference feature, a saturated water vapor pressure difference feature, a blood oxygen saturation level feature, a pulse rate feature and a blood oxygen consumption feature per unit time, and the infrared spectrum feature includes a dominant frequency energy feature and/or a dominant frequency feature.
In one embodiment, the obtaining the metabolic heat conformation characteristic of the blood glucose test object according to the physiological parameter and the environmental parameter comprises: obtaining temperature difference characteristics according to the body surface temperature and the environment temperature; and/or acquiring a first water vapor pressure corresponding to the body surface temperature and a second water vapor pressure corresponding to the environment temperature, and acquiring a water vapor pressure difference characteristic according to the first water vapor pressure, the second water vapor pressure and the environment humidity; and/or, using the blood oxygen saturation as the blood oxygen saturation characteristic; and/or obtaining pulse rate characteristics according to the multi-channel pulse signals; and/or obtaining the blood oxygen consumption characteristics in unit time according to the blood oxygen saturation, the multi-channel pulse signals and the electrocardiosignals. In one embodiment, the obtaining of the infrared spectrum characteristic of the blood glucose object according to the physiological parameter and the environmental parameter includes: taking the main frequency energy of the multi-channel pulse signal as a main frequency energy characteristic; and/or taking the main frequency of the multi-channel pulse signals as the main frequency characteristic.
And step 230, inputting the first input characteristic into the blood glucose measurement model to obtain the blood glucose value of the object to be measured through the blood glucose measurement model.
It can be understood that the method has the functions and the advantages corresponding to the parameter acquisition module, the characteristic determination module and the blood sugar determination module in the device capable of realizing non-invasive blood sugar detection, and technical details which are not described in the method can be described in the module corresponding to the parameters.
Fig. 3 is a flowchart of another method for performing non-invasive blood glucose measurement according to an embodiment of the present application, which can be performed by the apparatus for performing non-invasive blood glucose measurement. The method capable of realizing the noninvasive blood glucose detection is specifically a training process of a blood glucose measurement model. Referring to fig. 3, the method for enabling noninvasive blood glucose measurement includes:
and 310, acquiring a training characteristic and a target characteristic, wherein the training characteristic comprises a second input characteristic of the at least one object with known blood sugar, the second input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the object with known blood sugar, and the target characteristic comprises a blood sugar value of the at least one object with known blood sugar.
And step 320, constructing a blood glucose measurement model through the training characteristics and the target characteristics.
In one embodiment, when the blood glucose measurement model is constructed by partial least squares regression, step 320 specifically includes steps 321 to 328:
step 321, taking the training feature as a current first feature, and taking the target feature as a current second feature.
And 322, obtaining a first principal component axis vector of the current first feature according to the current first feature and the current second feature.
In one embodiment, when solving the first principal component axis vector and the second principal component axis vector by using a lagrange multiplier method, the step 322 specifically includes steps 3221 to 3222:
step 3221, mapping the current first feature to the first candidate principal component axis vector to obtain a first candidate principal component of the current first feature, and mapping the current second feature to the second candidate principal component axis vector to obtain a second candidate principal component of the current second feature.
Step 3222, determining a first candidate principal component axis vector and a second candidate principal component axis vector when the first candidate principal component and the second candidate principal component are optimal by using a lagrange multiplier method, and taking the first candidate principal component axis vector when the first candidate principal component and the second candidate principal component are optimal as a first principal component axis vector of the current first feature.
Step 323, determining a first principal component of the current first feature and a first residual error required when the first principal component and the first principal component axis vector represent the current first feature according to the first principal component axis vector.
In one embodiment, step 323 specifically includes steps 3231-3232:
step 3231, mapping the current first feature to the first principal component axis vector to obtain a first principal component of the current first feature.
Step 3232, determining a first residual error required when the first principal component and the first principal component axis vector represent the current first feature according to the first principal component and the first principal component axis vector.
And step 324, determining a target coefficient and a second residual error required when the current second characteristic is represented by the first principal component according to the first principal component and the current second characteristic.
In one embodiment, the target coefficient is calculated by a least squares method and the second residual is calculated by the target coefficient. In this case, step 324 specifically includes steps 3241 to 3243:
step 3241, a regression expression equation of the current second feature is constructed by using the first principal component, the target coefficient and the second residual error.
Step 3242, determine the target coefficients in the regression expression equation by least squares.
Step 3243, substituting the first principal component and the target coefficient into a regression expression equation to obtain a second residual error.
Step 325, determine whether the iteration stop condition is satisfied. If the iteration stop condition is not satisfied, go to step 326. If the iteration stop condition is satisfied, step 327 is executed.
The iteration stop condition includes that the current iteration number reaches a target iteration number.
Step 326, update the first residual to the current first feature and the second residual to the current second feature, and return to step 322.
It can be understood that, in practical applications, it is not necessary to set an operation of updating the current first feature and the current second feature, but it is only necessary to calculate the principal components of the first residual and the second residual according to steps 322 to 325 and obtain a new first residual and a new second residual.
And 327, constructing a linear expression of the target feature with respect to the training feature by using the first principal component axis vector, the first principal component and the target coefficient obtained in each iteration.
In one embodiment, step 327 may specifically include steps 3271-3272:
step 3271, a first expression of the target feature is constructed by using the first principal component and the target coefficient obtained by each iteration.
And step 3272, replacing the first principal component in the first expression by the training feature and the first principal component axis vector obtained by each iteration to obtain a linear expression of the target feature relative to the training feature.
And 328, taking the linear expression as a regression equation used by the blood glucose measurement model to complete the construction of the blood glucose measurement model.
It can be understood that the method has the functions and the advantages corresponding to the characteristic acquisition module and the model construction module in the device capable of realizing the non-invasive blood glucose detection, and technical details which are not described in the method can be described in the module corresponding to the parameters.
Fig. 4 is a schematic structural diagram of an apparatus capable of performing non-invasive blood glucose detection according to an embodiment of the present disclosure. As shown in fig. 4, the apparatus capable of noninvasive blood glucose detection includes a processor 40, a memory 41, an input device 42, and an output device 43; the number of processors 40 in the device capable of non-invasive blood glucose detection may be one or more, and one processor 40 is illustrated in fig. 4. The processor 40, the memory 41, the input device 42 and the output device 43 of the apparatus for performing non-invasive blood glucose detection may be connected by a bus or other means, and fig. 4 illustrates the connection by the bus as an example.
The memory 41 serves as a computer-readable storage medium, and can be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules (e.g., parameter obtaining module, feature determining module, blood glucose determining module, etc.) corresponding to the apparatus capable of non-invasive blood glucose detection in the embodiments of the present application. The processor 40 executes various functional applications of the apparatus that enables non-invasive blood glucose detection and data processing by executing software programs, instructions and modules stored in the memory 41.
The memory 41 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created from use of the apparatus that enables noninvasive blood glucose detection, and the like. Further, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 41 may further include memory located remotely from processor 40, which may be connected via a network to a device that may enable non-invasive blood glucose testing. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 42 may be used to receive input numeric or character information and generate key signal inputs related to user settings and functional control of the apparatus enabling non-invasive blood glucose testing, such as inputting blood glucose values of users with known blood glucose levels, and may also include triggers for collecting environmental and physiological parameters. The output device 43 may include a display device such as a display screen, which can display the measured blood glucose level and the like.
The equipment capable of realizing the noninvasive blood glucose detection comprises a device capable of realizing the noninvasive blood glucose detection, and has corresponding functions and beneficial effects.
In addition, the present application also provides a storage medium containing computer executable instructions, which when executed by a computer processor, are configured to perform relevant operations in the method for non-invasive blood glucose detection provided in any of the embodiments of the present application, and have corresponding functions and advantages.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product.
Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. 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 flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams 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. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory. The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.
Computer-readable media, including both permanent and non-permanent, removable and non-removable media, may implement the information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include transitory computer readable media (transmyedia) such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional identical elements in any process, method, article, or apparatus that comprises the element.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present application and the technical principles employed. Those skilled in the art will appreciate that the present application is not limited to the particular embodiments described herein, but is capable of many obvious modifications, rearrangements and substitutions without departing from the scope of the application. Therefore, although the present application has been described in more detail with reference to the above embodiments, the present application is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present application, and the scope of the present application is determined by the scope of the appended claims.
Claims (10)
- An apparatus for enabling non-invasive blood glucose testing, comprising: the device comprises a parameter acquisition module, a characteristic determination module and a blood sugar determination module;the parameter acquisition module is used for acquiring physiological parameters of a blood sugar object to be detected and environmental parameters of an area where the blood sugar object to be detected is located currently, and the physiological parameters are acquired in a non-invasive mode;the characteristic determination module is used for obtaining a first input characteristic of the blood sugar object to be tested according to the physiological parameter and the environmental parameter, wherein the first input characteristic comprises an infrared spectrum characteristic and a metabolic heat integration characteristic of the blood sugar object to be tested;the blood sugar determining module is used for inputting the first input characteristics into a blood sugar measuring model so as to obtain the blood sugar value of the object to be measured for blood sugar through the blood sugar measuring model.
- The apparatus capable of performing non-invasive blood glucose detection according to claim 1, further comprising: a characteristic obtaining module and a model constructing module,the characteristic acquisition module is used for acquiring training characteristics and target characteristics, the training characteristics comprise second input characteristics of at least one object with known blood sugar, the second input characteristics comprise infrared spectrum characteristics and metabolic heat integration characteristics of the object with known blood sugar, and the target characteristics comprise blood sugar values of the at least one object with known blood sugar;the model building module is used for building the blood sugar measurement model through the training features and the target features.
- The apparatus capable of noninvasive blood glucose detection of claim 2, wherein the model building module comprises:a current feature determining unit, configured to use the training feature as a current first feature, and use the target feature as a current second feature;a first principal component axis vector determining unit, configured to obtain a first principal component axis vector of the current first feature according to the current first feature and the current second feature;a first residual determining unit configured to determine, according to the first principal component axis vector, a first principal component of the current first feature and a first residual required when the first principal component and the first principal component axis vector represent the current first feature;a second residual determining unit, configured to determine, according to the first principal component and the current second feature, a target coefficient and a second residual that are required when the current second feature is represented by the first principal component;the updating unit is used for updating the first residual error into a current first characteristic, updating the second residual error into a current second characteristic, and returning to execute the operation of obtaining a first principal component axis vector of the current first characteristic according to the current first characteristic and the current second characteristic until an iteration stop condition is met;the linear expression construction unit is used for constructing a linear expression of the target feature relative to the training feature by using the first principal component axis vector, the first principal component and the target coefficient obtained by each iteration;and the model construction unit is used for taking the linear expression as a regression equation used by the blood glucose measurement model so as to finish construction of the blood glucose measurement model.
- The apparatus enabling noninvasive blood glucose detection of claim 3, wherein the first principal component axis vector determination unit comprises:a mapping subunit, configured to map the current first feature to a first candidate principal component axis vector to obtain a first candidate principal component of the current first feature, and map the current second feature to a second candidate principal component axis vector to obtain a second candidate principal component of the current second feature;and an axis vector determining subunit, configured to determine, by using a lagrange multiplier method, a first candidate principal component axis vector and a second candidate principal component axis vector when the first candidate principal component and the second candidate principal component are optimal, and use the first candidate principal component axis vector when the first candidate principal component and the second candidate principal component are optimal as the first principal component axis vector of the current first feature.
- The apparatus capable of implementing a non-invasive blood glucose detecting method according to claim 3, wherein the first residual determining unit comprises:a first principal component determining subunit, configured to map the current first feature to the first principal component axis vector, so as to obtain a first principal component of the current first feature;a first residual calculating subunit, configured to determine, according to the first principal component and the first principal component axis vector, a first residual required when the first principal component and the first principal component axis vector represent the current first feature.
- The apparatus enabling non-invasive blood glucose testing according to claim 3, wherein the second residual error determining unit includes:a regression expression equation constructing subunit, configured to construct a regression expression equation of the current second feature using the first principal component, the target coefficient, and the second residual error;a target coefficient determination subunit configured to determine the target coefficient in the regression expression equation by a least square method;and the second residual error calculation subunit is configured to substitute the first principal component and the target coefficient into the regression expression equation to obtain the second residual error.
- The apparatus for enabling noninvasive glucose sensing of claim 3, wherein the iteration stop condition comprises a current number of iterations reaching a target number of iterations.
- The apparatus capable of achieving noninvasive blood glucose detection of claim 3, wherein the linear expression construction unit comprises:the first expression constructing subunit is used for constructing a first expression of the target feature by using the first principal component and the target coefficient obtained by each iteration;and the linear expression determining subunit is used for replacing the first principal component in the first expression by using the training feature and the first principal component axis vector obtained by each iteration to obtain a linear expression of the target feature relative to the training feature.
- The apparatus capable of performing non-invasive blood glucose detection according to claim 1, wherein the environmental parameters comprise an environmental temperature and an environmental humidity, and the physiological parameters comprise a body surface temperature, a blood oxygen saturation, an electrocardiograph signal and a multichannel pulse signal;the feature determination module includes:the metabolic heat integration characteristic determining unit is used for obtaining a temperature difference characteristic according to the body surface temperature and the environment temperature; and/or acquiring a first water vapor pressure corresponding to the body surface temperature and a second water vapor pressure corresponding to the environment temperature, and acquiring a water vapor pressure difference characteristic according to the first water vapor pressure, the second water vapor pressure and the environment humidity; and/or, using the blood oxygen saturation as a blood oxygen saturation characteristic; and/or obtaining pulse rate characteristics according to the multi-channel pulse signals; and/or obtaining blood oxygen consumption characteristics in unit time according to the blood oxygen saturation, the multichannel pulse signals and the electrocardiosignals;the infrared spectrum characteristic determining unit is used for taking the main frequency energy of the multi-channel pulse signal as a main frequency energy characteristic; and/or, taking the main frequency of the multi-channel pulse signal as a main frequency characteristic.
- An apparatus for performing non-invasive blood glucose measurement, comprising the apparatus for performing non-invasive blood glucose measurement according to any one of claims 1-9.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2021/100746 WO2022261917A1 (en) | 2021-06-17 | 2021-06-17 | Apparatus that can implement non-invasive blood glucose testing, and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115768348A true CN115768348A (en) | 2023-03-07 |
Family
ID=84526108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202180024619.4A Pending CN115768348A (en) | 2021-06-17 | 2021-06-17 | Device and equipment capable of realizing noninvasive blood glucose detection |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN115768348A (en) |
| WO (1) | WO2022261917A1 (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103349553A (en) * | 2013-08-01 | 2013-10-16 | 赵巍 | Double-wavelength differential near-infrared non-invasive glucose meter |
| CN106980746A (en) * | 2016-12-16 | 2017-07-25 | 清华大学 | A kind of general Woundless blood sugar Forecasting Methodology based on Time-Series analysis |
| CN107174259A (en) * | 2017-06-26 | 2017-09-19 | 上海理工大学 | Woundless blood sugar value harvester and computational methods based on multi-wavelength conservation of energy |
| CN107174258A (en) * | 2017-06-02 | 2017-09-19 | 北京信息科技大学 | Blood sugar concentration Forecasting Methodology |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2152895A2 (en) * | 2007-05-11 | 2010-02-17 | Sigmed, Inc. | Non-invasive characterization of a physiological parameter |
| CN105962949A (en) * | 2016-06-14 | 2016-09-28 | 上海理工大学 | Noninvasive blood glucose calculating method based on near-infrared light energy conservation law and signal collecting device |
| CN110575181A (en) * | 2019-09-10 | 2019-12-17 | 重庆大学 | Near infrared spectrum noninvasive blood glucose detection network model training method |
-
2021
- 2021-06-17 CN CN202180024619.4A patent/CN115768348A/en active Pending
- 2021-06-17 WO PCT/CN2021/100746 patent/WO2022261917A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103349553A (en) * | 2013-08-01 | 2013-10-16 | 赵巍 | Double-wavelength differential near-infrared non-invasive glucose meter |
| CN106980746A (en) * | 2016-12-16 | 2017-07-25 | 清华大学 | A kind of general Woundless blood sugar Forecasting Methodology based on Time-Series analysis |
| CN107174258A (en) * | 2017-06-02 | 2017-09-19 | 北京信息科技大学 | Blood sugar concentration Forecasting Methodology |
| CN107174259A (en) * | 2017-06-26 | 2017-09-19 | 上海理工大学 | Woundless blood sugar value harvester and computational methods based on multi-wavelength conservation of energy |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022261917A1 (en) | 2022-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Habbu et al. | Estimation of blood glucose by non-invasive method using photoplethysmography | |
| US20160256114A1 (en) | Non-invasive blood sugar measuring method and fingertip measuring probe | |
| Natarajan et al. | Photoplethysmography fast upstroke time intervals can be useful features for cuff-less measurement of blood pressure changes in humans | |
| CN110575181A (en) | Near infrared spectrum noninvasive blood glucose detection network model training method | |
| Chen et al. | Machine learning method for continuous noninvasive blood pressure detection based on random forest | |
| CN110279404A (en) | A kind of pulse wave acquisition device and method | |
| CN112133442A (en) | Continuous non-invasive blood glucose detection device and method | |
| Zhang et al. | Non-invasive blood glucose detection system based on conservation of energy method | |
| US20230172565A1 (en) | Systems, devices, and methods for developing a model for use when performing oximetry and/or pulse oximetry and systems, devices, and methods for using a fetal oximetry model to determine a fetal oximetry value | |
| WO2020221265A1 (en) | Noninvasive human blood component concentration monitoring method and apparatus | |
| JP2012085877A (en) | Concentration determination apparatus, concentration determination method, and program | |
| CN115768348A (en) | Device and equipment capable of realizing noninvasive blood glucose detection | |
| CN114786576A (en) | System and method for performing trans-abdominal fetal oximetry or pulse oximetry | |
| Altunkan et al. | A comparative study of an ambulatory blood pressure measuring device and a wrist blood pressure monitor with a position sensor versus a mercury sphygmomanometer | |
| Morales et al. | Respiratory rate estimation on embedded system | |
| RU2804226C1 (en) | Wearable device to determine the central body temperature of the user, method for determining the central body temperature of the user using the wearable device and system comprising the wearable device | |
| Melnyk et al. | Investigation of Wireless pulse oximeters for smartphone-based remote monitoring of lung health | |
| da Silva | A pervasive system for real-time blood pressure monitoring | |
| Du et al. | Adaptive network-based fuzzy inference system for assessment of lower limb peripheral vascular occlusive disease | |
| Wang et al. | Photoplethysmography-based heart action monitoring using a growing multilayer network | |
| Chen et al. | A blood pressure estimation approach based on single-channel photoplethysmography differential features | |
| WO2022270098A1 (en) | Blood neutral fat estimation device, blood neutral fat estimation method, and program | |
| Bikia et al. | Cardiac output estimated from an uncalibrated radial blood pressure waveform: validation in an in-silico-generated population | |
| US20230142121A1 (en) | Fusion of sensor data for persistent disease monitoring | |
| CN113631088B (en) | Calibration-free pulse oximetry |
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 |