DEVICE AND METHOD FOR OBTAINING AN INDIVIDUALIZED UNIT SPECTRUM AND DEVICE AND METHOD FOR ESTIMATING A BIOLOGICAL COMPONENT

Description

BACKGROUND1. Field

[0001] Apparatuses and methods consistent with exemplary embodiments relate to estimating biological components in a non-invasive manner, and more particularly to obtaining an individualized unit spectrum in consideration of optical characteristics of individuals, and estimating biological components based on the individualized unit spectrum.2. Description of the Related Art

[0002] Diabetes is a chronic disease that causes various complications and can be hardly cured, such that people with diabetes are advised to check their blood glucose regularly to prevent complications. In particular, when insulin is administered to control blood glucose, the blood glucose levels have to be closely monitored to avoid hypoglycemia and control insulin dosage. An invasive method of finger pricking is generally used to measure blood glucose levels. However, while the invasive method may provide high reliability in measurement, it may cause pain and inconvenience as well as an increased risk of disease infections due to the use of a lancet for the finger pricking. Recently, research has been conducted on non-invasive measurements of blood glucose by using a spectrometer without collecting blood.

[0003] An example in the art is provided by US 2012 / 035442 A1 that relates to the selection of the specific filter combinations, which can provide sufficient information for multivariate calibration to extract accurate analyte concentrations in complex biological systems. This document also describes wavelength interval selection methods that give rise to the miniaturized designs. Finally, it presents a plurality of wavelength selection methods and miniaturized spectroscopic apparatus designs and the necessary tools to map from one domain (wavelength selection) to the other (design parameters). Such selection of informative spectral bands has a broad scope in miniaturizing any clinical diagnostic instruments which employ Raman spectroscopy in particular and other spectroscopic techniques in general.

[0004] Another example is given by WO 2004 / 081524 A2 that provides a solution for reducing interference in noninvasive spectroscopic measurements of tissue and blood analytes. By applying a basis set representing various tissue components to a collected sample measurement, measurement interferences resulting from the heterogeneity of tissue, sampling site differences, patient-to-patient variation, physiological variation, and instrumental differences are reduced. Consequently, the transformed sample measurements are more suitable for developing calibrations that are robust with respect to sample-to-sample variation, variation through time, and instrument related differences. In the calibration phase, data associated with a particular tissue sample site is corrected using a selected subset of data within the same data set. This method reduces the complexity of the data and reduces the intra-subject, inter-subject, and inter-instrument variations by removing interference specific to the respective data subset. In the measurement phase, the basis set correction is applied using a minimal number of initial samples collected from the sample site(s) where future samples will be collected.SUMMARY

[0005] The apparatus according to independent claim 1 and the corresponding method according to independent claim 9 present the present invention, while embodiments of the invention are presented in the dependent claims. The invention and its embodiments address at least the above problems and / or disadvantages and other disadvantages not described above. However, it is not required to overcome the disadvantages described above, and it is not required to overcome any of the problems described above.

[0006] The present invention and its embodiments provide an apparatus and method for estimating a biological component.

[0007] According to an aspect of the present invention, there is obtained an individualized unit spectrum by obtaining a first biological spectrum from a subject at a first measurement time, and obtaining a second biological spectrum from the subject at a second measurement time. Additionally, there is a processor configured to extract the individualized unit spectrum from the first biological spectrum and the second biological spectrum, based on a predetermined unit spectrum of a target component.

[0008] The processor includes a candidate spectrum extractor configured to extract at least one candidate spectrum from the first biological spectrum and the second biological spectrum.

[0009] The candidate spectrum extractor further is configured to extract the at least one candidate spectrum based on at least one of principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), and auto-encoding (AE).

[0010] The processor further includes: a candidate spectrum selector configured to select a candidate spectrum, associated with the target component, from among the extracted at least one candidate spectrum; and an individualized unit spectrum extractor configured to scale the selected candidate spectrum based on the predetermined unit spectrum, and extract the scaled candidate spectrum as the individualized unit spectrum.

[0011] The candidate spectrum selector is further configured to select the candidate spectrum based on at least one of a shape of the candidate spectrum and a variation over time in the candidate spectrum.

[0012] The candidate spectrum selector is further configured to select the candidate spectrum in response to the candidate spectrum satisfying at least one of a first condition under which a first similarity between the candidate spectrum and the predetermined unit spectrum exceeds a first threshold, a second condition under which a second similarity between a variation over time in the candidate spectrum and a step function exceeds a second threshold, and a third condition under which a product of the first similarity and the second similarity exceeds a third threshold.

[0013] The candidate spectrum selector uses a similarity calculation method or a statistical test, wherein the similarity calculation method includes at least one of Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, and Spearman's Correlation Coefficient, and the statistical test may include at least one of f-test, t-test, and z-test.

[0014] In response to a plurality of candidate spectrums being selected by the candidate spectrum selector, the individualized unit spectrum extractor is configured to generate an average candidate spectrum by averaging the selected plurality of candidate spectrums, and scale the generated average candidate spectrum to correspond to the predetermined unit spectrum.

[0015] The apparatus further includes a preprocessor configured to remove noise from the first biological spectrum and the second biological spectrum.

[0016] The preprocessor is further configured to use at least one of asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), and Savitzky-Golay smoothing (SG).

[0017] The processor also includes: a subtractor configured to subtract the first biological spectrum from the second biological spectrum to obtain a subtracted spectrum; and an individualized unit spectrum extractor configured to scale the subtracted spectrum to correspond to the predetermined unit spectrum of the target component, and extract the scaled subtracted spectrum as the individualized unit spectrum.

[0018] The processor is further configured to select a method from among: a first method of extracting the individualized unit spectrum using principal component analysis (PCA); a second method of extracting the individualized unit spectrum using independent component analysis (ICA); a third method of extracting the individualized unit spectrum using non-negative matrix factorization (NMF); a fourth method of extracting the individualized unit spectrum using auto-encoding (AE); and a fifth option of method the individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum, wherein the processor may be further configured to extract the individualized unit spectrum based on the selected method.

[0019] The target component includes a blood component and a skin component of the subject, wherein the blood component includes at least one of blood glucose, cholesterol, triglyceride, proteins, and uric acid; and the skin component includes at least one of proteins including collagen, keratin, and elastin, and body fat.

[0020] According to another aspect of the present invention, there is provided a corresponding method in which an individualized unit spectrum is obtained by: obtaining a first biological spectrum from a subject at a first measurement time, and obtaining a second biological spectrum from the subject at a second measurement time; and extracting the individualized unit spectrum from the first biological spectrum and the second biological spectrum, based on a predetermined unit spectrum of a target component.

[0021] The extracting the individualized unit spectrum includes: extracting at least one candidate spectrum from the first biological spectrum and the second biological spectrum; selecting a candidate spectrum, associated with the target component, from the extracted at least one candidate spectrum; scaling the selected candidate spectrum based on the predetermined unit spectrum; and extracting the scaled candidate spectrum as the individualized unit spectrum.

[0022] The extracting the candidate spectrum includes extracting the candidate spectrum based on at least one of principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), and auto-encoding (AE).

[0023] The selecting the candidate spectrum associated with the target component includes selecting the candidate spectrum based on at least one of a shape of the candidate spectrum and a variation over time in the candidate spectrum.

[0024] The selecting the candidate spectrum associated with the target component includes selecting the candidate spectrum in response to the candidate spectrum satisfying at least one of a first condition under which a first similarity between the candidate spectrum and the predetermined unit spectrum exceeds a first threshold, a second condition under which a second similarity between a variation over time in the candidate spectrum and a step function exceeds a second threshold, and a third condition under which a product of the first similarity and the second similarity exceeds a third threshold.

[0025] The selecting the candidate spectrum associated with the target component includes selecting the candidate spectrum based on a similarity calculation method or a statistical test, wherein the similarity calculation method may include at least one of Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, and Spearman's Correlation Coefficient, and the statistical test may include at least one of f-test, t-test, and z-test.

[0026] The scaling the selected candidate spectrum includes, in response to a plurality of candidate spectrums being selected, generating an average candidate spectrum by averaging the selected plurality of candidate spectrums, and scaling the generated average candidate spectrum to correspond to the predetermined unit spectrum.

[0027] The method further includes removing noise from the first biological spectrum and the second biological spectrum.

[0028] The removing the noise includes removing the noise based on at least one of asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), and Savitzky-Golay smoothing (SG).

[0029] The extracting the individualized unit spectrum includes: subtracting the first biological spectrum from the second biological spectrum to obtain a subtracted spectrum; scaling the subtracted spectrum to correspond to the predetermined unit spectrum of the target component; and extracting the scaled subtracted spectrum as the individualized unit spectrum.

[0030] The method further includes selecting a method from among: a first method of extracting the individualized unit spectrum using principal component analysis (PCA); a second method of extracting the individualized unit spectrum using independent component analysis (ICA); a third method of extracting the individualized unit spectrum using non-negative matrix factorization (NMF); a fourth method of extracting the individualized unit spectrum using auto-encoding (AE); and a fifth method of extracting the individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum, wherein the extracting the individualized unit spectrum comprises extracting the individualized unit spectrum based on the selected method.

[0031] The target component includes a blood component and a skin component, wherein the blood component includes at least one of blood glucose, cholesterol, triglyceride, proteins, and uric acid; and the skin component may include at least one of proteins including collagen, keratin, and elastin, and body fat.

[0032] The apparatus according to present invention further comprises a second processor thatincludes: a model generator configured to generate a biological component estimation model based on a background spectrum of the user and the individualized unit spectrum; and a biological component estimator configured to estimate the biological component based on the measured biological spectrum and the biological component estimation model.BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The above and / or other aspects will be more apparent by describing certain exemplary embodiments, with reference to the accompanying drawings, in which: FIG. 1 is a block diagram illustrating an apparatus for obtaining an individualized unit spectrum according to an exemplary embodiment; FIG. 2 is a block diagram illustrating a processor according to an exemplary embodiment; FIG. 3 is a block diagram illustrating a processor according to another exemplary embodiment; FIG. 4 is a block diagram illustrating an apparatus for obtaining an individualized unit spectrum according to another exemplary embodiment; FIG. 5 illustrates a diagram explaining extracting an individualized unit spectrum by using principal component analysis (PCA) according to an exemplary embodiment; FIG. 6 is a diagram explaining extracting an individualized unit spectrum by using auto-encoding (AE) according to an exemplary embodiment; FIG. 7 is a block diagram illustrating an apparatus for estimating a biological component according to an exemplary embodiment; FIG. 8 is a block diagram illustrating a spectrum measuring apparatus according to an exemplary embodiment; FIG. 9 is a block diagram illustrating a processor according to another exemplary embodiment; FIG. 10 is a block diagram illustrating an apparatus for estimating a biological component according to another exemplary embodiment; FIG. 11 is a block diagram illustrating an apparatus for estimating a biological component according to another exemplary embodiment; FIG. 12 is a flowchart illustrating a method of obtaining an individualized unit spectrum according to an exemplary embodiment; FIG. 13 is a block diagram illustrating a method of extracting an individualized unit spectrum according to an exemplary embodiment; FIG. 14 is a block diagram illustrating a method of extracting an individualized unit spectrum according to another exemplary embodiment; FIG. 15 is a flowchart illustrating a method of obtaining an individualized unit spectrum according to another exemplary embodiment; and FIG. 16 is a flowchart illustrating a method of estimating a biological component according to an exemplary embodiment. DETAILED DESCRIPTION

[0034] Exemplary embodiments are described in greater detail below with reference to the accompanying drawings.

[0035] In the following description, like drawing reference numerals are used for like elements, even in different drawings. The matters defined in the description, such as detailed construction and elements, are provided to assist in a comprehensive understanding of the exemplary embodiments. However, it is apparent that the exemplary embodiments can be practiced without those specifically defined matters. Also, well-known functions or constructions are not described in detail since they would obscure the description with unnecessary detail.

[0036] Process steps described herein may be performed differently from a specified order, unless a specified order is clearly stated in the context of the disclosure. That is, each step may be performed in a specified order, at substantially the same time, or in a reverse order.

[0037] Further, the terms used throughout this specification are defined in consideration of the functions according to exemplary embodiments, and can be varied according to a purpose of a user or manager, or precedent and so on. Therefore, definitions of the terms should be made on the basis of the overall context.

[0038] It will be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. Any references to singular may include plural unless expressly stated otherwise. In the present specification, it should be understood that the terms, such as 'including' or 'having,' etc., are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof disclosed in the specification, and are not intended to preclude the possibility that one or more other features, numbers, steps, actions, components, parts, or combinations thereof may exist or may be added.

[0039] Expressions such as "at least one of," when preceding a list of elements, modify the entire list of elements and do not modify the individual elements of the list. For example, the expression, "at least one of a, b, and c," should be understood as including only a, only b, only c, both a and b, both a and c, both b and c, or all of a, b, and c.

[0040] Further, components that will be described in the specification are discriminated merely according to functions mainly performed by the components. That is, two or more components which will be described later can be integrated into a single component. Furthermore, a single component which will be explained later can be separated into two or more components. Moreover, each component which will be described can additionally perform some or all of a function executed by another component in addition to the main function thereof. Some or all of the main function of each component which will be explained can be carried out by another component. Each component may be implemented as hardware, software, or a combination of both.

[0041] A unit spectrum described in the present disclosure refers to a spectrum of a material per unit concentration (e.g., 1 mM), and an individualized unit spectrum refers to a unit spectrum obtained in consideration of optical characteristics of each individual.

[0042] FIG. 1 is a block diagram illustrating an apparatus for obtaining an individualized unit spectrum according to an exemplary embodiment. The apparatus 100 may be embedded in an electronic device. In particular, examples of the electronic device may include a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like; and examples of the wearable device may include a wristwatch-type wearable device, a wristband-type wearable device, a ring-type wearable device, a waist belt-type wearable device, a necklace-type wearable device, an ankle band-type wearable device, a thigh band-type wearable device, a forearm band-type wearable device, and the like. However, the electronic device is not limited thereto, and the wearable device is neither limited thereto.

[0043] Referring to FIG. 1, the apparatus 100 for obtaining an individualized unit spectrum includes a spectrum obtainer 110 and a processor 120.

[0044] The spectrum obtainer 110 obtains a biological spectrum (hereinafter referred to as a first biological spectrum) measured in the case where there is a small amount of a target component in a body, and a biological spectrum (hereinafter referred to as a second biological spectrum) measured in the case where there is a large amount of a target component in a body. The target component includes a blood component, including blood glucose, cholesterol, triglyceride, protein, uric acid, and the like, and a skin component including collagen, keratin, elastin, and the like.

[0045] In one exemplary embodiment, the spectrum obtainer 110 obtains the first biological spectrum and the second biological spectrum from an external device which measures and / or stores biological spectrums. In particular, the spectrum obtainer 110 uses various communication techniques, such as Bluetooth communication, Bluetooth Low Energy (BLE) communication, Near Field Communication (NFC), WLAN communication, Zigbee communication, Infrared Data Association (IrDA) communication, Wi-Fi Direct (WFD) communication, Ultra-Wideband (UWB) communication, Ant+ communication, WiFi communication, Radio Frequency Identification (RFID) communication, 3G communication, 4G communication, 5G communication, and the like.

[0046] In another exemplary embodiment, in the case where there is a small amount of a target component in a body (e.g., in an empty stomach state if a target component is blood glucose), the spectrum obtainer 110 obtains the first biological spectrum by emitting light onto skin and by receiving light reflected or scattered from the skin; and in the case where there is a large amount of a target component (e.g., after intake of sugar if a target component is blood glucose), the spectrum obtainer 110 obtains the second biological spectrum by emitting light onto skin and by receiving light reflected or scattered from the skin. For example, the spectrum obtainer 100 obtains the first biological spectrum at a first measurement time (e.g., 5 hours after food consumption), and obtains the second biological spectrum at a second measurement time (e.g., within 2 hours after food consumption). The spectrum obtainer 110 includes a light source to emit light onto skin, and a photodetector to obtain a biological spectrum by receiving light reflected or scattered from the skin. For example, the spectrum obtainer 110 is realized as an optical spectrometer.

[0047] The light source in the spectrum obtainer 110 emits near infrared rays (NIR) or mid infrared rays (MIR). However, wavelengths of light to be emitted by the light source may vary according to a purpose of measurement or the types of target component to be analyzed. Further, the light sourcemay include a single light-emitting body, or may be formed as an array of a plurality of light-emitting bodies. The light source may include a light emitting diode (LED), a laser diode, a fluorescent body, and the like. The photodetector may include a photo diode, a photo transistor (PTr), a charge-coupled device (CCD), and the like. The photodetector may include a single device, and may be formed as an array of a plurality of devices. There may be various numbers and arrangements of light sources and photodetectors, and the number and arrangement thereof may vary according to the types of target component, a purpose of use, the size and shape of the electronic device in which the apparatus 100 is embedded, and the like.

[0048] The processor 120 processes various signals and operations related to the obtaining of an individualized unit spectrum by the apparatus 100.

[0049] According to predetermined intervals or a user's request, the processor 120 controls the spectrum obtainer 110 to obtain the first biological spectrum and the second biological spectrum, and extracts an individualized unit spectrum, corresponding to a unit spectrum of a target component, from the obtained first biological spectrum and second biological spectrum. Here, various intervals may be set by a user to repeatedly obtain the first biological spectrum and the second biological spectrum. The unit spectrum is experimentally derived in advance and stored in the apparatus 100.

[0050] For example, the processor 120 extracts an individualized unit spectrum from the first biological spectrum or the second biological spectrum by using a feature extraction method, or extracts an individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum. In this case, the feature extraction method includes principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), auto-encoding (AE), and the like.

[0051] FIG. 2 is a block diagram illustrating a processor according to an exemplary embodiment. The processor 200 of FIG. 2 may be an example of the processor 120 of FIG. 1.

[0052] Referring to FIG. 2, the processor 200 includes a candidate spectrum extractor 210, a candidate spectrum selector 220, and an individualized unit spectrum extractor 230.

[0053] The candidate spectrum extractor 210 extracts at least one candidate spectrum from the first biological spectrum and the second biological spectrum by using a feature extraction method. In this case, as described above, the feature extraction method includes principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), auto-encoding (AE), and the like, as described above.

[0054] The candidate spectrum selector 220 selects a candidate spectrum, associated with a target component, from among the extracted at least one candidate spectrum. In one exemplary embodiment, the candidate spectrum selector 220 selects a candidate spectrum, which satisfies predetermined requirements, as a candidate spectrum associated with a target component based on the shape of a candidate spectrum and / or a variation over time in a candidate spectrum. In particular, the predetermined requirements includes a requirement for a similarity (hereinafter referred to as a first similarity) between a candidate spectrum and a unit spectrum of a target component to exceed a predetermined threshold (hereinafter referred to as a first threshold), a requirement for a similarity (hereinafter referred to as a second similarity) between a variation over time in a candidate spectrum and a step function to exceed a predetermined threshold (hereinafter referred to as a second threshold), a requirement for a value obtained by multiplying the first similarity and the second similarity to exceed a predetermined threshold (hereinafter referred to as a third threshold), and the like. The step function schematically represents an actual change in the target component, and information on a unit spectrum of a target component is pre-stored in an internal or external database. The candidate spectrum selector 220 selects a candidate spectrum, which satisfies at least one of the aforementioned predetermined requirements, from among the extracted at least one candidate spectrum.

[0055] For example, the candidate spectrum selector 220 uses a similarity calculation method, such as Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, Spearman's Correlation Coefficient, and the like, or a statistical test such as f-test, t-test, z-test, and the like.

[0056] The individualized unit spectrum extractor 230 scales the selected candidate spectrum to correspond to a unit spectrum of a target component, and extracts the scaled candidate spectrum as an individualized unit spectrum. The candidate spectrum is extracted from the first biological spectrum and the second biological spectrum regardless of a concentration of a target component, such that the individualized unit spectrum extractor 230 extracts an individualized unit spectrum, corresponding to a unit spectrum, by scaling the selected candidate spectrum to a range of a unit spectrum. The scaled selected candidate spectrum has substantially the same amplitude as the unit spectrum of the target component.

[0057] In addition, the candidate spectrum selector 220 selects a plurality of candidate spectrums. In this case, the individualized unit spectrum extractor 230 generates an average candidate spectrum by averaging the selected plurality of candidate spectrums, scales the generated average candidate spectrum to correspond to a unit spectrum, and extracts the scaled average candidate spectrum as an individualized unit spectrum.

[0058] FIG. 3 is a block diagram illustrating a processor according to another exemplary embodiment. The processor 300 of FIG. 3 may be another example of the processor 120 of FIG. 1.

[0059] Referring to FIG. 3, the processor 300 includes a subtractor 310 and an individualized unit spectrum extractor 320.

[0060] The subtractor 310 subtracts the first biological spectrum from the second biological spectrum. In one exemplary embodiment, in the case where there are a plurality of second biological spectrums and a plurality of first biological spectrums, the subtractor 310 calculates an average second biological spectrum and an average first biological spectrum, and subtracts the calculated average first biological spectrum from the calculated average second biological spectrum.

[0061] The individualized unit spectrum extractor 320 scales the spectrum, generated as a result of the subtraction by the subtractor 310, to correspond to a unit spectrum of a target component, and extracts the scaled spectrum as the individualized unit spectrum.

[0062] FIG. 4 is a block diagram illustrating an apparatus for obtaining an individualized unit spectrum according to another exemplary embodiment. The apparatus 400 is embedded in an electronic device. In particular, examples of the electronic device may include a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like; and examples of the wearable device may include a wristwatch-type wearable device, a wristband-type wearable device, a ring-type wearable device, a waist belt-type wearable device, a necklace-type wearable device, an ankle band-type wearable device, a thigh band-type wearable device, a forearm band-type wearable device, and the like. However, the electronic device is not limited thereto, and the wearable device is neither limited thereto.

[0063] Referring to FIG. 4, the apparatus 400 further includes selectively a preprocessor 410, an option selector 420, an input interface 430, a storage 440, a communication interface 450, and an output interface 460, in addition to a spectrum obtainer 110 and a processor 120. Here, the spectrum obtainer 110 and the processor 120 are described above with reference to FIGS. 1 to 3, such that detailed description thereof will be omitted. Further, FIG. 4 illustrates the preprocessor 410 and the option selector 420 as separate parts from the processor 120, this is merely exemplary for convenience of explanation, and the preprocessor 410 and the option selector 420 may be configured as a part of the processor 120.

[0064] The preprocessor 410 removes noise, which occurs by components other than a target component, from the first biological spectrum and the second biological spectrum. In one exemplary embodiment, the preprocessor 310 uses various noise removal methods, such as asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), Savitzky-Golay smoothing (SG), and the like, which are merely exemplary, and the noise removal method is not limited thereto.

[0065] The option selector 420 selects one of a plurality of options related to a method of extracting an individualized unit spectrum. In particular, the plurality of options include a first option of extracting an individualized unit spectrum by using principal component analysis (PCA), a second option of extracting an individualized unit spectrum by using independent component analysis (ICA), a third option of extracting an individualized unit spectrum by using non-negative matrix factorization (NMF), a fourth option of extracting an individualized unit spectrum by using auto-encoding (AE), and a fifth option of extracting an individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum. In this case, the processor 120 extracts an individualized unit spectrum from the first biological spectrum and the second biological spectrum by using at least one of the plurality of options.

[0066] The input interface 430 receives input of various operation signals from a user. In one exemplary embodiment, the input interface 430 may include a keypad, a dome switch, a touch pad (static pressure / capacitance), a jog wheel, a jog switch, a hardware (H / W) button, and the like. Particularly, the touch pad, which forms a layer structure with a display, may be called a touch screen.

[0067] The storage 440 stores programs or commands for operation of the apparatus 400, and stores data input to and output from the apparatus 400. Further, the storage 440 stores the first biological spectrum data and the second biological spectrum data which are obtained by the spectrum obtainer 110, an individualized unit spectrum data extracted by the processor 120, a unit spectrum data of a target component, and the like.

[0068] The storage 440 includes at least one storage medium of a flash memory type memory, a hard disk type memory, a multimedia card micro type memory, a card type memory (e.g., an SD memory, an XD memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, and an optical disk, and the like. Further, the apparatus 400 operates an external storage medium, such as web storage and the like, which performs a storage function of the storage 440 on the Internet.

[0069] The communication interface 450 performs communication with an external device. For example, the communication interface 450 transmits, to the external device, data input by a user through the input interface 430, the first biological spectrum data and the second biological spectrum data which are obtained by the spectrum obtainer 110, the individualized unit spectrum data extracted by the processor 120, the unit spectrum data of a target component, and the like, or receives, from the external device, various data useful for extracting an individualized unit spectrum.

[0070] In this case, the external device is medical equipment using the first biological spectrum data and the second biological spectrum data which are obtained by the spectrum obtainer 110, the individualized unit spectrum data extracted by the processor 120, the unit spectrum data of a target component, and the like, a printer to print out results, or a display to display the extracted individualized unit spectrum data. In addition, the external device may be a digital TV, a desktop computer, a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like, but is not limited thereto.

[0071] The communication interface 450 communicates with an external device by using Bluetooth communication, Bluetooth Low Energy (BLE) communication, Near Field Communication (NFC), WLAN communication, Zigbee communication, Infrared Data Association (IrDA) communication, Wi-Fi Direct (WFD) communication, Ultra-Wideband (UWB) communication, Ant+ communication, WIFI communication, Radio Frequency Identification (RFID) communication, 3G communication, 4G communication, 5G communication, and the like. However, this is merely exemplary and is not intended to be limiting.

[0072] The output interface 460 outputs the first biological spectrum data and the second biological spectrum data which are obtained by the spectrum obtainer 110, the individualized unit spectrum data extracted by the processor 120, the unit spectrum data of a target component, and the like. In one exemplary embodiment, the output interface 460 outputs the first biological spectrum data and the second biological spectrum data which are obtained by the spectrum obtainer 110, the individualized unit spectrum data extracted by the processor 120, the unit spectrum data of a target component, and the like by using at least one of an acoustic method, a visual method, and a tactile method. To this end, the output interface 460 includes a display, a speaker, a vibrator, and the like.

[0073] FIG. 5 illustrates a diagram explaining extracting an individualized unit spectrum by using principal component analysis (PCA) according to an exemplary embodiment. For convenience of explanation, it is assumed in FIG. 5 that one candidate spectrum is extracted.

[0074] Referring to FIGS. 2 and 5, the candidate spectrum extractor 210 extracts, as a candidate spectrum, an eigenvector 510 of a first principal component based on the first biological spectrum and the second biological spectrum.

[0075] The candidate spectrum selector 220 determines a similarity R1 between the eigenvector 510 of the first principal component and a unit spectrum 530 of a target component, and determines whether the similarity R1 exceeds the first threshold. The unit spectrum 530 is experimentally derived in advance and stored in the storage 440. Further, the candidate spectrum selector 220 determines a similarity R2 between the eigenvalue 520 of the first principal component and a step function 540, and determines whether the similarity R2 exceeds the second threshold. The processor 120 uses the step function 540 to approximate or estimate a glucose level change in reality. It may be obtained by measuring a change in blood glucose concentrations over time in a step function.

[0076] In the case where the similarity R1 exceeds the first threshold, and the similarity R2 exceeds the second threshold, the candidate spectrum selector 220 selects the eigenvector 510 of the first principal component as a candidate spectrum associated with a target component.

[0077] The individualized unit spectrum extractor 230 scales the eigenvector 510 of the first principal component to correspond to the unit spectrum 530, and determines the scaled eigenvector as an individualized unit spectrum. In order to make the eigenvector 510 correspond to the unit spectrum 530, the individualized unit spectrum extractor 230 scales the eigenvector 510 by adjusting an amplitude of the eigenvector 510 to match the amplitude of the unit spectrum 530. The scaled eigenvector 510 have substantially the same waveform as the original eigenvector 510, but may have a different amplitude from the original eigenvector 510.

[0078] Although it is assumed in FIG. 5 that one candidate spectrum is extracted, this is merely exemplary for convenience of explanation, and the extraction is not limited to one candidate spectrum. That is, the candidate spectrum extractor 210 extracts eigenvectors of the first principal component to an n-th principal component, which are generated by using PCA, as candidate spectrums; and the candidate spectrum selector 220 selects one eigenvector as a candidate spectrum associated with a target component by determining similarity between each of the eigenvectors of the first principal component to the n-th principal component and the unit spectrum 530, and by determining similarity between each of the eigenvectors of the first principal component to the n-th principal component and the step function 540. In this case, 'n' may be set to various values according to performance and purpose of use of a system.

[0079] The eigenvector may be referred to as a loading vector, a latent variable, a principal component, and the like, and the eigenvalue may be referred to as a loading score and the like.

[0080] FIG. 6 is a diagram explaining extracting an individualized unit spectrum by using auto-encoding (AE) according to an exemplary embodiment. For convenience of explanation, it is assumed in FIG. 6 that one candidate spectrum is extracted.

[0081] Referring to FIGS. 2 and 6, the candidate spectrum extractor 210 extracts a weight vector 610 as a candidate spectrum based on the first biological spectrum and the second biological spectrum, by using AE such as, for example, denoising auto-encoding, sparse auto-encoding, variational auto-encoding, and contractive auto-encoding.

[0082] The candidate spectrum selector 220 determines a similarity R3 between the weight vector 610 and a unit spectrum 630 of a target component, and determines whether the similarity R3 exceeds a third threshold. The unit spectrum 630 is experimentally derived in advance and has the same value as the unit spectrum 530. Further, the candidate spectrum selector 220 determines a similarity R4 between a value 620, obtained by multiplying an input data and the weight vector 610, and a step function 640, and determines whether the similarity R4 exceeds a fourth threshold. The input data may refer to data which are inputted to perform the AE, such as the first biological spectrum and the second biological spectrum. The step function 640 may have substantially the same function as the step function 540.

[0083] In the case where the similarity R3 exceeds the third threshold, and the similarity R4 exceeds the fourth threshold, the candidate spectrum selector 220 selects the weight vector 610 as a candidate spectrum associated with a target component.

[0084] The individualized unit spectrum extractor 230 scales the weight vector 610 to correspond to the unit spectrum 630, and extracts the scaled weight vector as an individualized unit spectrum.

[0085] Although it is assumed in FIG. 6 that one candidate spectrum is extracted, this is merely exemplary for convenience of explanation, and the extraction is not limited to one candidate spectrum. That is, similarly to FIG. 5, the candidate spectrum extractor 210 may extract, as a candidate spectrum, a plurality of weight vectors by using AE.

[0086] FIG. 7 is a block diagram illustrating an apparatus for estimating a biological component according to exemplary embodiment. The biological component estimating apparatus 700 may be embedded in an electronic device. In particular, examples of the electronic device may include a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like; and examples of the wearable device may include a wristwatch-type wearable device, a wristband-type wearable device, a ring-type wearable device, a waist belt-type wearable device, a necklace-type wearable device, an ankle band-type wearable device, a thigh band-type wearable device, a forearm band-type wearable device, and the like. However, the electronic device is not limited thereto, and the wearable device is neither limited thereto.

[0087] Referring to FIG. 7, the biological component estimating apparatus 700 includes a spectrum measurer 710 and a processor 720.

[0088] The spectrum measurer 710 measures a biological spectrum from skin of a user. The spectrum measurer 710 measures the biological spectrum by emitting light onto the skin and by receiving light reflected or scattered from the skin according to a predetermined control signal. In one exemplary embodiment, the spectrum measurer 710 uses infrared spectroscopy or Raman spectroscopy, but is not limited thereto, and measures a spectrum by using various spectroscopy techniques.

[0089] The processor 720 processes various operations associated with estimation of a user's biological component by analyzing the measured biological spectrum data. Here, the biological component includes a blood component, including blood glucose, cholesterol, triglyceride, proteins, uric acid, and the like, and a skin component including collagen, keratin, elastin, and the like.

[0090] Hereinafter, for convenience of explanation, the following description will be made based on an example where a biological component is blood glucose.

[0091] The processor 720 estimates a user's blood glucose value based on the measured biological spectrum and an individualized unit spectrum. In particular, a method of obtaining the individualized unit spectrum is described above with reference to FIGS. 1 to 6, such that detailed description thereof will be omitted.

[0092] In one exemplary embodiment, the processor 720 estimates a blood glucose value by using the following Equations 1 and 2. BS = Sm − Sb BG = RG + BS Su

[0093] Herein, Sm denotes the measured biological spectrum, Sb denotes a background spectrum, BS denotes a biological spectrum from which the background spectrum is removed, Su denotes the individualized unit spectrum, RG denotes a reference blood glucose value, and BG denotes the estimated blood glucose value. In this case, the background spectrum Sb may be spectrums measured continuously at regular intervals for a predetermined period of time in a reference state (e.g., when there is a small amount of blood glucose in a body, i.e., in an empty stomach state), and the reference state may be defined differently for each user according to user characteristics and the like. The reference blood glucose value RG may refer to a blood glucose value in a fasting state.

[0094] That is, once the biological spectrum Sm is measured for estimating a blood glucose value, the processor 720 may remove noise, caused by a component other than blood glucose, by subtracting the background spectrum Sb from the biological spectrum Sm, in order to increase a signal to noise ratio (SNR). Further, the processor 720 may estimate a blood glucose variation value BS Su based on the biological spectrum BS , from which noise (background spectrum) is removed, and the individualized unit spectrum Su , and may calculate the estimated blood glucose value by adding the reference blood glucose value RG to the estimated blood glucose variation value BS Su .

[0095] As the biological component estimating apparatus 700 estimates a biological component by using an individualized unit spectrum in consideration of optical characteristics of each spectrum, the biological component estimating apparatus 700 estimates a biological component of a user more accurately.

[0096] FIG. 8 is a block diagram illustrating a spectrum measuring apparatus according to an exemplary embodiment. The spectrum measuring apparatus 800 may be an example of the spectrum measurer 710 of FIG. 7.

[0097] Referring to FIG. 8, the spectrum measuring apparatus 800 includes a spectrometer 810 and a measurement controller 820.

[0098] The spectrometer 810 includes a light source 811 to emit light onto a user's skin, and a photodetector 812 to obtain a skin spectrum by receiving light reflected or scattered from the skin.

[0099] The light source 811 emits near infrared rays (NIR) or mid infrared rays (MIR) onto the user's skin. However, wavelengths of light to be emitted by the light source 811 may vary according to a purpose of measurement or the types of target component to be analyzed. Further, the light source 811 includes a single light-emitting body, or may be formed as an array of a plurality of light-emitting bodies. The light source 811 may be a light emitting diode (LED), a laser diode, a fluorescent body, and the like.

[0100] The photodetector 812 includes a photo diode, a photo transistor (PTr), a charge-coupled device (CCD), and the like. The photodetector 812 includes a single device, or may be formed as an array of a plurality of devices.

[0101] There may be various numbers and arrangements of the light source 811 and the photodetector 812, and the number and arrangement thereof may vary according to the types of target component, a purpose of use, the size and shape of the electronic device in which the spectrum measuring apparatus 800 is embedded, and the like.

[0102] The user's skin, onto which light is emitted, may be an area on a wrist that is adjacent to the radial artery. In the case where the area is the skin surface of the wrist where the radial artery passes, measurement may be relatively less affected by external factors, such as the thickness of a skin tissue in the wrist, which may cause errors in measurement. However, the object is not limited thereto, and may be a peripheral body part of the human body, such as fingers, toes, and the like, where blood vessels are densely located.

[0103] The measurement controller 820 controls the spectrometer 810 by generating a control signal according to a user's command or predetermined criteria.

[0104] In one exemplary embodiment, the measurement controller 820 is connected to a healthcare instrument (e.g., a probe) in which the spectrum measuring apparatus 800 is mounted, and generates a control signal to control the spectrometer 810 by receiving a command for measuring a spectrum which is received through the instrument. In this case, the command for measuring a spectrum is a command for measuring a biological spectrum for estimating a biological component, or a command for measuring a background spectrum for use in noise removal.

[0105] In another exemplary embodiment, measurement criteria of a biological spectrum for estimating a biological component or a background spectrum is preset, and the measurement controller 820 may automatically generate a control signal for controlling the spectrometer 810 according to the set criteria. For example, a time or an interval of estimating blood glucose levels is preset for users, such as diabetic patients, for whom it is highly important to manage their blood glucose level, so that blood glucose levels are estimated at a regular time or at regular intervals in an empty stomach state or after intake of sugar. Further, a measurement interval of a background spectrum, which is measured to remove noise from the measured biological spectrum, is preset.

[0106] The measurement controller 820 is described as a part of the spectrum measuring apparatus 800, but is not limited thereto. That is, the measurement controller 820 may be configured as a part of the processor 720 of FIG. 7.

[0107] FIG. 9 is a block diagram illustrating a processor according to an exemplary embodiment. The processor 900 of FIG. 9 may be an example of the processor 720 of FIG. 7.

[0108] Referring to FIG. 9, the processor 900 includes a model generator 910 and a biological component estimator 920.

[0109] The model generator 910 generates a biological component estimation model for use in estimating a biological component. The model generator 910 generates a biological component estimation model or updates an existing biological component estimation model by using a background spectrum Sb and an individualized unit spectrum Su.

[0110] In one exemplary embodiment, the model generator 910 generates a blood glucose estimation model, represented by the following Equation 3, by using Beer Lambert's Law. However, the blood glucose estimation model of Equation 3 is merely exemplary, and the model is not limited thereto. Sm = Sb 1 + Sb 2 + Sb 3 + Sb 4 + ⋯ + ε g ⋅ L ⋅ Cg

[0111] Herein, Sm denotes a user's biological spectrum measured for estimating blood glucose, (Sb1 + Sb2 + Sb3 + Sb4 + ···) denotes background spectrums measured for a predetermined period of time, ε g denotes an individualized unit spectrum for blood glucose, L denotes a light travel length while the biological spectrum Sm is measured or a light travel length estimated by using an aqueous path length, and Cg denotes a blood glucose variation value of a user.

[0112] The biological component estimator 920 estimates a biological component of a user by using the measured biological spectrum and the biological component estimation model.

[0113] In one exemplary embodiment, the biological component estimator 920 calculates the blood glucose variation value Cg of a user by substituting the biological spectrum Sm in Equation 3, and calculates an estimated blood glucose value by adding a reference blood glucose value to the blood glucose variation value Cg (see Equation 2).

[0114] FIG. 10 is a block diagram illustrating an apparatus for estimating a biological component according to another exemplary embodiment. The biological component estimating apparatus 1000 is embedded in an electronic device. In particular, examples of the electronic device may include a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like; and examples of the wearable device may include a wristwatch-type wearable device, a wristband-type wearable device, a ring-type wearable device, a waist belt-type wearable device, a necklace-type wearable device, an ankle band-type wearable device, a thigh band-type wearable device, a forearm band-type wearable device, and the like. However, the electronic device is not limited thereto, and the wearable device is neither limited thereto.

[0115] Referring to FIG. 10, the biological component estimating apparatus 1000 includes a spectrum measurer 1010, a first processor 1020, and a second processor 1030.

[0116] The spectrum measurer 1010 performs the function of the spectrum obtainer 110 of FIG. 1 and the function of the spectrum measurer 710 of FIG. 7, the first processor 1020 performs the function of the processor 120 of FIG. 1, and the second processor 1030 performs the function of the processor 720 of FIG 7. That is, the apparatus 1000 for estimating a biological component is configured by integrating the apparatus 100 for obtaining an individualized unit spectrum of FIG. 1 and the apparatus 700 for estimating a biological component of FIG. 7.

[0117] Although FIG. 10 illustrates the first processor 1020 and the second processor 1030 as separate parts, the first processor 1020 and the second processor 1030 may be integrated into a single processor.

[0118] FIG. 11 is a block diagram illustrating yet another example of an apparatus for estimating a biological component.

[0119] Referring to FIG. 11, the biological component estimating apparatus 1100 includes a spectrum measurer 710, a processor 720, an input interface 1110, a storage 1120, a communication interface 1130, and an output interface 1140. Here, the spectrum measurer 710 and the processor 720 are described above with reference to FIGS. 7 to 9, such that detailed description thereof will be omitted.

[0120] The input interface 1110 receives input of various operation signals from a user. In one exemplary embodiment, the input interface 1110 may include a keypad, a dome switch, a touch pad (static pressure / capacitance), a jog wheel, a jog switch, a hardware (H / W) button, and the like. Particularly, the touch pad, which forms a layer structure with a display, may be called a touch screen.

[0121] The storage 1120 stores programs or commands for operation of the apparatus 1100 for estimating a biological component, and stores data input to and output from the apparatus 1100 for estimating a biological component. Further, the storage 1120 stores biological spectrum data measured by the spectrum measurer 710, estimated biological component data estimated by the processor 720, individualized unit spectrum data of a biological component, background spectrum data, biological component estimation model data, and the like.

[0122] The storage 1120 includes at least one storage medium of a flash memory type memory, a hard disk type memory, a multimedia card micro type memory, a card type memory (e.g., an SD memory, an XD memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, and an optical disk, and the like. Further, the apparatus 1100 for estimating a biological component operates an external storage medium, such as web storage and the like, which performs a storage function of the storage 1120 on the Internet.

[0123] The communication interface 1130 performs communication with an external device. For example, the communication interface 1130 transmits, to the external device, data input by a user through the input interface 1110, the biological spectrum data measured by the spectrum measurer 710, the estimated biological component data estimated by the processor 720, the individualized unit spectrum data of a biological component, the background spectrum data, the biological component estimation model data, and the like, or receives, from the external device, various data useful for estimating a biological component.

[0124] In particular, the external device is medical equipment using the biological spectrum data measured by the spectrum measurer 710, the estimated biological component data estimated by the processor 720, the individualized unit spectrum data of a biological component, the background spectrum data, the biological component estimation model data, and the like, a printer to print out results, or a display to display the results. In addition, the external device may be a digital TV, a desktop computer, a cellular phone, a smartphone, a tablet PC, a laptop computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation, an MP3 player, a digital camera, a wearable device, and the like, but is not limited thereto.

[0125] The communication interface 1130 communicates with an external device by using Bluetooth communication, Bluetooth Low Energy (BLE) communication, Near Field Communication (NFC), WLAN communication, Zigbee communication, Infrared Data Association (IrDA) communication, Wi-Fi Direct (WFD) communication, Ultra-Wideband (UWB) communication, Ant+ communication, WIFI communication, Radio Frequency Identification (RFID) communication, 3G communication, 4G communication, 5G communication, and the like. However, this is merely exemplary and is not intended to be limiting.

[0126] The output interface 1140 outputs the biological spectrum data measured by the spectrum measurer 710, the estimated biological component data estimated by the processor 720, the individualized unit spectrum data of a biological component, the background spectrum data, the biological component estimation model data, and the like. In one exemplary embodiment, the output interface 1140 outputs the biological spectrum data measured by the spectrum measurer 710, the estimated biological component data estimated by the processor 720, the individualized unit spectrum data of a biological component, the background spectrum data, the biological component estimation model data, and the like by using at least one of an acoustic method, a visual method, and a tactile method. To this end, the output interface 1140 includes a display, a speaker, a vibrator, and the like.

[0127] FIG. 12 is a flowchart illustrating a method of obtaining an individualized unit spectrum according to an exemplary embodiment. The method of obtaining an individualized unit spectrum of FIG. 12 is performed by the apparatus 100 for obtaining an individualized unit spectrum of FIG. 1.

[0128] Referring to FIGS. 1 and 12, the apparatus 100 for obtaining an individualized unit spectrum obtains a first biological spectrum, which is measured when there is a small amount of a target component in a body, and a second biological spectrum which is measured when there is a large amount of a target component in a body in operation 1210.

[0129] For example, the apparatus 100 obtains the first biological spectrum and the second biological spectrum from an external device which measures and / or stores biological spectrums. Alternatively, the apparatus 100 obtains the first biological spectrum and the second biological spectrum by emitting light onto skin and by receiving light reflected or scattered from the skin. In particular, the apparatus 100 measures the first biological spectrum when the user's stomach is empty (e.g., two hours after the user consumes food), and measures the second biological spectrum when the user's stomach is full or almost full (e.g., within 30 minutes after the user consumes food).

[0130] The apparatus 100 extracts an individualized unit spectrum, corresponding to a unit spectrum of a target component, from the obtained first biological spectrum and second biological spectrum in operation 1220. For example, the processor 120 extracts the individualized unit spectrum from the first biological spectrum and the second biological spectrum by using a feature extraction method, or extracts an individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum.

[0131] FIG. 13 is a block diagram illustrating a method of extracting an individualized unit spectrum according to an exemplary embodiment. The method of extracting an individualized unit spectrum is an example of the extraction of an individualized unit spectrum in operation 1220 of FIG. 12.

[0132] Referring to FIGS. 1 and 13, the apparatus 100 extracts at least one candidate spectrum from the first biological spectrum and the second biological spectrum by using a feature extraction method in operation 1310. In this case, as described above, the feature extraction method includes principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), auto-encoding (AE), and the like.

[0133] The apparatus 100 selects a candidate spectrum, associated with a target component, from among the extracted at least one candidate spectrum in operation 1320. In one exemplary embodiment, the apparatus 100 selects, as a candidate spectrum associated with a target component, a candidate spectrum which satisfies predetermined requirements based on the shape of a candidate spectrum and / or a variation over time in a candidate spectrum. In particular, the predetermined requirements may include a requirement for a similarity (hereinafter referred to as first similarity) between a candidate spectrum and a unit spectrum of a target component to exceed a predetermined threshold (hereinafter referred to as a first threshold), a similarity (hereinafter referred to as second similarity) between a variation over time in a candidate spectrum and a step function to exceed a predetermined threshold (hereinafter referred to as a second threshold), a requirement of a value obtained by multiplying the first similarity and the second similarity to exceed a predetermined threshold (hereinafter referred to as a third threshold), and the like.

[0134] In particular, the apparatus 100 uses a similarity calculation method, such as Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, Spearman's Correlation Coefficient, and the like, or a statistical test such as f-test, t-test, z-test, and the like.

[0135] The apparatus 100 scales the selected candidate spectrum to correspond to a unit spectrum of a target component in operation 1330, and extracts the scaled candidate spectrum as an individualized unit spectrum in operation 1340.

[0136] Upon selecting a plurality of candidate spectrums in operation 1320, the apparatus 100 generates an average candidate spectrum by averaging the selected plurality of candidate spectrums, and scales the generated average candidate spectrum to correspond to a unit spectrum.

[0137] FIG. 14 is a block diagram illustrating a method of extracting an individualized unit spectrum according to another exemplary embodiment. The method of extracting an individualized unit spectrum of FIG. 14 is another example of the extraction of an individualized unit spectrum in operation 1220 of FIG. 12.

[0138] Referring to FIGS. 1 and 14, the apparatus 100 subtracts the first biological spectrum from the second biological spectrum in operation 1410. In one exemplary embodiment, in the case where there are a plurality of second biological spectrum and a plurality of first biological spectrum, the apparatus 100 calculates an average second biological spectrum and an average first biological spectrum, and subtracts the calculated average first biological spectrum from the calculated average second biological spectrum.

[0139] The apparatus 100, generated as a result of the subtraction in operation 1410, to correspond to a unit spectrum of a target component in operation 1420, and extracts the scaled spectrum as the individualized unit spectrum in operation 1430.

[0140] FIG. 15 is a flowchart illustrating a method of obtaining an individualized unit spectrum according to another exemplary embodiment. The method of obtaining an individualized unit spectrum of FIG. 15 is performed by the apparatus 400 for obtaining an individualized unit spectrum of FIG. 4.

[0141] Referring to FIGS. 4 and 15, the apparatus 400 obtains the first biological spectrum and the second biological spectrum in operation 1510, and removes noise, which occurs by a component other than a target component, from the first biological spectrum and the second biological spectrum in operation 1520. In one embodiment, the apparatus 400 uses various noise removal methods, such as asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), Savitzky-Golay smoothing (SG), and the like, which are merely exemplary, and the noise removal method is not limited thereto.

[0142] The apparatus 400 selects one of a plurality of options related to a method of extracting an individualized unit spectrum in 1530. In this case, the plurality of options include a first option of extracting an individualized unit spectrum by using principal component analysis (PCA), a second option of extracting an individualized unit spectrum by using independent component analysis (ICA), a third option of extracting an individualized unit spectrum by using non-negative matrix factorization (NMF), a fourth option of extracting an individualized unit spectrum by using auto-encoding (AE), and a fifth option of extracting an individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum.

[0143] The apparatus 400 extracts an individualized unit spectrum from the first biological spectrum and the second biological spectrum in operation 1540 by using a method of an option selected in operation 1530.

[0144] FIG. 16 is a flowchart illustrating an example of a method of estimating a biological component. The method of estimating a biological component of FIG. 16 is performed by the apparatus 700 for estimating a biological component of FIG. 7.

[0145] Referring to FIGS. 7 and 16, the biological component estimating apparatus 700 measures a biological spectrum from skin of a user. For example, the biological component estimating apparatus 700 measures a biological spectrum by emitting light onto the skin and by receiving light reflected or scattered from the skin in operation 1610.

[0146] The biological component estimating apparatus 700 estimates a blood glucose value of a user based on the measured spectrum and an individualized unit spectrum in operation 1620.

[0147] While not restricted thereto, an exemplary embodiment can be embodied as computer-readable code on a computer-readable recording medium. The computer-readable recording medium is any data storage device that can store data that can be thereafter read by a computer system. Examples of the computer-readable recording medium include read-only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices. The computer-readable recording medium can also be distributed over network-coupled computer systems so that the computer-readable code is stored and executed in a distributed fashion. Also, an exemplary embodiment may be written as a computer program transmitted over a computer-readable transmission medium, such as a carrier wave, and received and implemented in general-use or special-purpose digital computers that execute the programs. Moreover, it is understood that in exemplary embodiments, one or more units of the above-described apparatuses and devices can include circuitry, a processor, a microprocessor, etc., and may execute a computer program stored in a computer-readable medium.

Claims

1. An apparatus (1000) for estimating a biological component of a subject, the apparatus (1000) comprising: a spectrum measurer (1010; 110) configured to measure biological spectra by emitting light onto skin of the subject and by receiving light reflected or scattered from the skin of the subject, wherein the spectrum measurer (1010; 110) is further configured to obtain a first biological spectrum from the subject at a first measurement time, and obtain a second biological spectrum from the subject at a second measurement time when an amount of a target component in a body of the subject is larger than the amount at the first measurement time, wherein the first measurement time has a longer elapsed time than the second measurement time from a point in time when the subject consumes food, wherein the measured biological spectra including the first and second biological spectra are infrared or Raman spectra; a preprocessor (410) configured to remove noise from the first biological spectrum and the second biological spectrum; and a first processor (1020; 120; 200; 300) configured to extract an individualized unit spectrum from the first biological spectrum and the second biological spectrum, based on a predetermined unit spectrum of the target component; wherein the first processor (1020; 120; 200; 300) comprises: a subtractor (310) configured to subtract the first biological spectrum from the second biological spectrum to obtain a subtracted spectrum; a candidate spectrum extractor (210) configured to extract at least one candidate spectrum from the first biological spectrum and the second biological spectrum by using a feature extraction method; a candidate spectrum selector (220) configured to select a candidate spectrum, associated with the target component, from among the extracted at least one candidate spectrum based on satisfying at least one predetermined requirement based on a shape of the candidate spectrum and / or a variation over time in the candidate spectrum, wherein the at least one predetermined requirement includes a requirement for a first similarity between the candidate spectrum and the predetermined unit spectrum of the target component to exceed a first threshold, a requirement for a second similarity between a variation over time in the candidate spectrum and a step function representing an actual change in the target component to exceed a second threshold and a requirement for a value obtained by multiplying the first similarity and the second similarity to exceed a third threshold, and wherein the candidate spectrum selector (220) uses a similarity calculation method or a statistical test, wherein the similarity calculation method comprises at least one of Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, and Spearman's Correlation Coefficient, and wherein the statistical test comprises at least one of f-test, t-test, and z-test; and an individualized unit spectrum extractor (230; 320) configured to scale the selected candidate spectrum and / or the subtracted spectrum to correspond to the predetermined unit spectrum of the target component such that the scaled candidate spectrum and / or the scaled subtracted candidate spectrum have the same amplitude as the predetermined unit spectrum of the target component and extract the scaled candidate spectrum and / or the scaled subtracted spectrum as the individualized unit spectrum; and a second processor (1030; 720; 900) configured to estimate the biological component of the subject based on a measured biological spectrum and the individualized unit spectrum.

2. The apparatus of claim 1, wherein the candidate spectrum extractor is further configured to extract the at least one candidate spectrum based on at least one of the following feature extraction methods of principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), and auto-encoding (AE).

3. The apparatus of claim 1 or 2, wherein in response to a plurality of candidate spectrums being selected by the candidate spectrum selector (220), the individualized unit spectrum extractor (230; 320) is configured to generate an average candidate spectrum by averaging the selected plurality of candidate spectrums, and scale the generated average candidate spectrum to correspond to the predetermined unit spectrum.

4. The apparatus of one of the preceding claims, wherein the preprocessor (410) is further configured to use at least one of asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), and Savitzky-Golay smoothing (SG).

5. The apparatus of one of the preceding claims, wherein the first processor (1020; 120; 200; 300) is further configured to select a method from among: a first method of extracting the individualized unit spectrum using principal component analysis (PCA); a second method of extracting the individualized unit spectrum using independent component analysis (ICA); a third method of extracting the individualized unit spectrum using non-negative matrix factorization (NMF); a fourth method of extracting the individualized unit spectrum using auto-encoding (AE); and a fifth option of method the individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum, and wherein the first processor (1020; 120; 200; 300) is further configured to extract the individualized unit spectrum based on the selected method.

6. The apparatus of one of the preceding claims, wherein the target component comprises a blood component and a skin component of the subject, wherein the blood component comprises at least one of blood glucose, cholesterol, triglyceride, proteins, and uric acid; and the skin component comprises at least one of proteins including collagen, keratin, and elastin, and body fat.

7. The apparatus of one of the preceding claims, wherein the second processor (1030; 720; 900) further comprises: a model generator (910) configured to generate a biological component estimation model based on a background spectrum of the subject and the individualized unit spectrum of the target component; and a biological component estimator (920) configured to estimate the biological component based on the measured biological spectrum and the biological component estimation model.

8. The apparatus of claim 7, wherein the spectrum measurer (1010; 110) is further configured to measure the background spectrum from the subject at the first measurement time.

9. A computer-implemented method of estimating a biological component of a subject, wherein biological spectra have been measured by emitting light onto skin of the subject and by receiving light reflected or scattered from the skin of the subject, a first biological spectrum has been obtained from the subject at a first measurement time, and a second biological spectrum has been obtained from the subject at a second measurement time when an amount of a target component in a body of the subject is larger than the amount at the first measurement time, wherein the first measurement time has a longer elapsed time than the second measurement time from a point in time when the subject consumes food, and wherein the first and second biological spectra are infrared or Raman spectra, the method comprising: removing noise from the first biological spectrum and the second biological spectrum; and extracting an individualized unit spectrum from the first biological spectrum and the second biological spectrum, based on a predetermined unit spectrum of the target component, wherein the extracting the individualized unit spectrum comprises: subtracting the first biological spectrum from the second biological spectrum to obtain a subtracted spectrum; extracting at least one candidate spectrum from the first biological spectrum and the second biological spectrum using a feature extraction method; selecting a candidate spectrum, associated with the target component, from the extracted at least one candidate spectrum based on satisfying at least one predetermined requirement based on a shape of the candidate spectrum and / or a variation over time in the candidate spectrum, wherein the at least one predetermined requirement includes a requirement for a first similarity between the candidate spectrum and the predetermined unit spectrum of the target component to exceed a first threshold, a requirement for a second similarity between a variation over time in the candidate spectrum and a step function representing an actual change in the target component to exceed a second threshold and a requirement for a value obtained by multiplying the first similarity and the second similarity to exceed a third threshold, and wherein selecting the candidate spectrum associated with the target component comprises selecting the candidate spectrum based on a similarity calculation method or a statistical test, wherein the similarity calculation method comprises at least one of Euclidean distance, Manhattan Distance, Cosine Distance, Mahalanobis Distance, Jaccard Coefficient, Extended Jaccard Coefficient, Pearson's Correlation Coefficient, and Spearman's Correlation Coefficient, and wherein the statistical test comprises at least one of f-test, t-test, and z-test; scaling the selected candidate spectrum and / or the subtracted spectrum to correspond to the predetermined unit spectrum of the target component such that the scaled candidate spectrum and / or the scaled subtracted candidate spectrum have the same amplitude as the predetermined unit spectrum of the target component; extracting the scaled candidate spectrum and / or the scaled subtracted spectrum as the individualized unit spectrum, and wherein the extracting the scaled candidate spectrum comprises extracting the candidate spectrum based on at least one of principal component analysis (PCA), independent component analysis (ICA), non-negative matrix factorization (NMF), and auto-encoding (AE); and estimating the biological component of the subject based on a measured biological spectrum and the individualized unit spectrum.

10. The computer-implemented method of claim 9, wherein the scaling the selected candidate spectrum comprises, in response to a plurality of candidate spectrums being selected, generating an average candidate spectrum by averaging the selected plurality of candidate spectrums, and scaling the generated average candidate spectrum to correspond to the predetermined unit spectrum.

11. The computer-implemented method of one of the claims 9 or 10, further comprising removing the noise based on at least one of asymmetric least square (ALS), detrend, multiplicative scatter correction (MSC), extended multiplicative scatter correction (EMSC), standard normal variate (SNV), mean centering (MC), Fourier transform (FT), orthogonal signal correction (OSC), and Savitzky-Golay smoothing (SG).

12. The computer-implemented method of one of the claims 9 to 11, further comprising selecting a method from among: a first method of extracting the individualized unit spectrum using principal component analysis (PCA); a second method of extracting the individualized unit spectrum using independent component analysis (ICA); a third method of extracting the individualized unit spectrum using non-negative matrix factorization (NMF); a fourth method of extracting the individualized unit spectrum using auto-encoding (AE); and a fifth method of extracting the individualized unit spectrum by subtracting the first biological spectrum from the second biological spectrum, wherein the extracting the individualized unit spectrum comprises extracting the individualized unit spectrum based on the selected method.

13. The computer-implemented method of one of the claims 9 to 12, wherein the target component comprises a blood component and a skin component, wherein the blood component comprises at least one of blood glucose, cholesterol, triglyceride, proteins, and uric acid; and the skin component comprises at least one of proteins including collagen, keratin, and elastin, and body fat.

14. The computer-implemented method of one of claims 9 to 13, wherein estimating the biological component of the subject further comprises: generating a biological component estimation model based on a background spectrum of the subject and the individualized unit spectrum of the target component; and estimating the biological component based on the measured biological spectrum and the biological component estimation model.

15. The computer-implemented method of claim 14, wherein the background spectrum is measured from the subject at the first measurement time.