Method and device for non-contact oxygen saturation prediction based on hyperspectral camera

Abstract

A method for non-contact oxygen saturation prediction based on a hyperspectral camera, according to one embodiment of the present invention, may comprise the steps of: acquiring a hyperspectral image of a subject by using the hyperspectral camera; measuring the absorbance of at least one face region that is detected in consideration of the correlation with a change in oxygen saturation from among the face regions of the subject from the hyperspectral image; determining at least two optimal wavelength bands from among wavelength bands measured by using the hyperspectral camera in consideration of the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and predicting the oxygen saturation of the subject on the basis of the absorbance of the at least one face region and the at least two optimal wavelength bands.

Description

Non-contact oxygen saturation prediction method and device based on a hyperspectral camera

[0001] The present invention relates to a non-contact oxygen saturation prediction method and device based on a hyperspectral camera. This study was conducted with the support of the Korea Health Industry Development Institute (Project ID: 1465039863; Project ID: HW20C2077010123; Research and Development Project: Infectious Disease Prevention Technology Development; Research Project Title: Development of a Non-Contact Smart Vital Signs Measurement Device for Infectious Disease Patients; Project Period: 2023.01.01. ~ 2023.05.31.), and with the support of the National IT Industry Promotion Agency (NIPA) with the funding of the Ministry of Science and ICT (NIPA) (Project ID: 1711193399; Project ID: 2022-0-00067-002; Research and Development Project: Development of Smart Manufacturing Innovation Technology; Research Project Title: (Detail 4-2) Development of 5G Edge Brain-Based Intelligent Manufacturing Equipment and Robot Integrated Control Solution and Logistics Process Verification; Project Period: This research project was conducted with the support of the National IT Industry Promotion Agency (NIPA) funded by the Ministry of Science and ICT (Government) (Project ID: 1711193231; Project ID: 2019-0-00421-005; Research and Development Project: Training of Innovative Talents in Information and Communications Broadcasting; Research Project Name: Support for Graduate School of Artificial Intelligence (Sungkyunkwan University); Project Period: 2023.01.01. ~ 2023.12.31.), this research project was conducted with the support of the National IT Industry Promotion Agency (NIPA) funded by the Ministry of Science and ICT (Government) (Project ID: 1711193280; Project ID: 2020-0-01821-004; Research and Development Project: Training of Innovative Talents in Information and Communications Broadcasting; Research Project Name: Training of ICT Masterpiece Talents (Sungkyunkwan University); Project Period: (Related to 2023.01.01. ~ 2023.12.31.)

[0002] For reference, this application claims priority to Korean Patent Application No. 10-2024-0026391, filed on February 23, 2024. The entire contents of that application, which serves as the basis for this priority claim, are incorporated herein by reference.

[0003] Conventional contact-type pulse oximeters (fingertip oximeters) can cause measurement errors or impossibility of measurement depending on the position of the finger being measured. Furthermore, measuring oxygen saturation in infants and young children using conventional contact-type pulse oximeters is not only difficult, but the time it takes for arterialized blood to reach the measurement site is approximately 24 seconds, resulting in a measurement delay, making accurate oxygen saturation measurement difficult. To overcome this, research and attempts to measure oxygen saturation using non-contact methods are ongoing.

[0004] However, the conventional non-contact oxygen saturation measurement method uses surface reflectance and diffused reflectance measured using an RGB camera or a spectral camera. However, since surface reflectance is sensitively affected by external environmental factors, there is a disadvantage in that the measurement data values ​​are highly volatile.

[0005] Additionally, in the case of surface reflection reflected from the skin surface, there is a disadvantage in that it does not accurately reflect information on the dermis layer of the skin, which is closely related to changes in oxygen saturation.

[0006] Accordingly, a method for minimizing the influence of surface reflection needs to be developed to increase the accuracy of oxygen saturation prediction.

[0007] The problem to be solved by the present invention is to enable more accurate measurement of oxygen saturation by receiving only the diffuse reflection component from the inner side of the skin, using this to remove the influence of the blood flow component, and removing the shadow according to the structure of the human face.

[0008] However, the problems to be solved by the present invention are not limited to those mentioned above, and other problems to be solved that are not mentioned can be clearly understood by a person having ordinary skill in the art to which the present invention pertains from the description below.

[0009] A non-contact oxygen saturation prediction method based on a hyperspectral camera according to one embodiment of the present invention may include: obtaining a hyperspectral image of a subject using the hyperspectral camera; measuring absorbance of at least one facial region detected from the hyperspectral image in consideration of a correlation with a change in the oxygen saturation among the facial regions of the subject; determining at least two optimal wavelength bands among wavelength bands measured using the hyperspectral camera in consideration of the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and predicting the oxygen saturation of the subject based on the absorbance of the at least one facial region and the at least two optimal wavelength bands.

[0010] Here, the step of acquiring the hyperspectral image may include a step of acquiring a hyperspectral image including only a diffuse reflection component by using a first polarizing filter attached to the hyperspectral camera and a second polarizing filter attached to the light so as to be orthogonal to the direction of the first polarizing filter.

[0011] Additionally, the step of measuring the absorbance may include a step of removing shadow elements of the face area of ​​the subject by considering the geometric structure of the face area of ​​the subject.

[0012] Additionally, in the step of measuring the absorbance, at least one facial area detected in consideration of the correlation with the change in the oxygen saturation can be detected based on an image processing algorithm.

[0013] Here, the at least one facial region may include at least one of a forehead region, an under-eye region, a cheek region, and an under-nose region.

[0014] Meanwhile, the method may further include a step of removing a blood flow component based on a ratio of absorbance measured in at least two optimal wavelength bands determined above, and a step of predicting the oxygen saturation based on at least two optimal wavelength bands from which the blood flow component has been removed.

[0015] In addition, the step of predicting the oxygen saturation may include a step of predicting the oxygen saturation for each of the at least one facial region based on the absorbance measured time-seriesly for each of the at least one facial region and the at least two optimal wavelength bands; and a step of predicting the oxygen saturation for the subject using the oxygen saturation for each of the at least one facial region.

[0016] According to another embodiment of the present invention, a non-contact oxygen saturation prediction device based on a hyperspectral camera includes: a memory in which an oxygen saturation prediction program is stored; and a processor for loading the oxygen saturation prediction program from the memory and executing the oxygen saturation prediction program, wherein the processor obtains a hyperspectral image of a subject using the hyperspectral camera, measures absorbance of at least one facial region detected from the hyperspectral image in consideration of a correlation with a change in the oxygen saturation among the facial regions of the subject, and determines at least two optimal wavelength bands among wavelength bands measured using the hyperspectral camera in consideration of the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin, and predicts the oxygen saturation of the subject based on the absorbance of the at least one facial region and the at least two optimal wavelength bands.

[0017] Here, the processor can obtain a hyperspectral image including only a diffuse reflection component by using a first polarizing filter attached to the hyperspectral camera and a second polarizing filter attached to the light so as to be orthogonal to the direction of the first polarizing filter.

[0018] Additionally, the processor can remove shadow elements of the face area of ​​the subject by considering the geometric structure of the face area of ​​the subject.

[0019] Additionally, at least one facial region detected in consideration of its correlation with the change in oxygen saturation may be detected based on an image processing algorithm.

[0020] Here, the at least one facial region may include a forehead region, an under-eye region, a cheek region, and an under-nose region.

[0021] Meanwhile, the processor can remove the blood flow component based on the ratio of absorbance measured in at least two optimal wavelength bands determined above.

[0022] Additionally, the processor can predict oxygen saturation for each of the at least one facial region based on absorbance measured time-series for each of the at least one facial region and the at least two optimal wavelength bands, and can predict oxygen saturation for the subject using the oxygen saturation for each of the at least one facial region.

[0023] According to another embodiment of the present invention, a computer-readable recording medium storing a computer program may include instructions for causing the processor to perform a non-contact oxygen saturation prediction method based on a hyperspectral camera, the method comprising: obtaining a hyperspectral image of a subject using a hyperspectral camera; measuring an absorbance of at least one facial region detected from the hyperspectral image in consideration of a correlation with a change in the oxygen saturation among the facial regions of the subject; determining at least two optimal wavelength bands among wavelength bands measured using the hyperspectral camera in consideration of the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and predicting the oxygen saturation of the subject based on the absorbance of the at least one facial region and the at least two optimal wavelength bands.

[0024] According to another embodiment of the present invention, a computer program stored in a computer-readable recording medium may include instructions for causing the processor to perform a non-contact oxygen saturation prediction method based on a hyperspectral camera, the method comprising: obtaining a hyperspectral image of a subject using a hyperspectral camera; measuring an absorbance of at least one facial region detected from the hyperspectral image in consideration of a correlation with a change in the oxygen saturation among the facial regions of the subject; determining at least two optimal wavelength bands among wavelength bands measured using the hyperspectral camera in consideration of the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and predicting the oxygen saturation of the subject based on the absorbance of the at least one facial region and the at least two optimal wavelength bands.

[0025] According to an embodiment of the present invention, the possibility of infection of medical staff and the general public can be minimized by non-contact measurement of the patient's oxygen saturation status.

[0026] In addition, according to an embodiment of the present invention, by using a polarizing filter attached to a hyperspectral camera and a polarizing filter attached to a light so as to be orthogonal to the direction of the polarizing filter, the influence of surface reflection is eliminated, and a hyperspectral image containing only diffuse reflection components is obtained, thereby enabling more precise measurement of oxygen saturation.

[0027] In addition, according to an embodiment of the present invention, oxygen saturation can be measured more precisely by eliminating differences in absorbance that occur depending on the different facial shapes of each person.

[0028] Fig. 1 is a block diagram showing an oxygen saturation prediction device according to an embodiment of the present invention.

[0029] FIG. 2 is a block diagram conceptually illustrating the function of an oxygen saturation prediction program according to an embodiment of the present invention.

[0030] Figure 3 is a flowchart illustrating a method for predicting oxygen saturation according to one embodiment of the present invention.

[0031] FIG. 4 is a drawing exemplarily showing obtaining only a diffuse reflection component using a polarizing filter attached to a hyperspectral camera and lighting according to one embodiment of the present invention.

[0032] FIG. 5 is a drawing exemplarily showing the removal of shadow elements from a face area of ​​a subject according to one embodiment of the present invention.

[0033] FIG. 6 is a drawing exemplarily showing determining at least two wavelength bands for predicting oxygen saturation by considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin according to one embodiment of the present invention.

[0034] The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided solely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined solely by the scope of the claims.

[0035] When describing embodiments of the present invention, detailed descriptions of known functions or configurations will be omitted if they are deemed to unnecessarily obscure the gist of the invention. Furthermore, the terms described below are defined in light of their functions in the embodiments of the present invention and may vary depending on the intent or custom of the user or operator. Therefore, their definitions should be based on the overall content of this specification.

[0036] Fig. 1 is a block diagram showing an oxygen saturation prediction device according to an embodiment of the present invention.

[0037] Referring to FIG. 1, the oxygen saturation prediction device (100) may include a processor (110), an input / output device (120), and a memory (130).

[0038] The processor (110) can control the overall operation of the oxygen saturation prediction device (100).

[0039] The processor (110) can obtain a hyperspectral image of the subject using the input / output device (120).

[0040] Specifically, the processor (110) can acquire hyperspectral images of a subject in a time series manner using a hyperspectral camera.

[0041] In the present invention, a hyperspectral image may mean an image having spectral bands of hundreds of continuous spectral wavelengths reflected from a subject or object.

[0042] In the present invention, the hyperspectral image is described as being input through the input / output device (120), but is not limited thereto. That is, depending on the embodiment, the oxygen saturation prediction device (100) may include a transceiver (not shown), the oxygen saturation prediction device (100) may receive a hyperspectral image using the transceiver (not shown), and the hyperspectral image may be generated within the oxygen saturation prediction device (100).

[0043] The processor (110) acquires a hyperspectral image of a subject using a hyperspectral camera, measures absorbance of at least one facial region detected from the hyperspectral image in consideration of correlation with oxygen saturation among facial regions of the subject, determines at least two wavelength bands among wavelength bands measured using the hyperspectral camera in consideration of absorbance of oxyhemoglobin and absorbance of deoxyhemoglobin, and predicts oxygen saturation of the subject based on the absorbance of at least one facial region and the at least two wavelength bands.

[0044] The input / output device (120) may include one or more input devices and / or one or more output devices. For example, the input devices may include a microphone, a keyboard, a mouse, a touch screen, etc., and the output devices may include a display, a speaker, etc.

[0045] The memory (130) can store the oxygen saturation prediction program (200) and information necessary for executing the oxygen saturation prediction program (200).

[0046] In this specification, the oxygen saturation prediction program (200) may mean software including commands for predicting the oxygen saturation of the subject by receiving a hyperspectral image of the subject captured using a hyperspectral camera.

[0047] The processor (110) can load the oxygen saturation prediction program (200) and information necessary for executing the oxygen saturation prediction program (200) from the memory (130) to execute the oxygen saturation prediction program (200).

[0048] The processor (110) can execute an oxygen saturation prediction program (200) to receive a hyperspectral image and measure the absorbance of at least one facial region detected by considering the correlation with the change in oxygen saturation in the facial region of the subject.

[0049] Additionally, the processor (110) can determine at least two wavelength bands among the wavelength bands measured using the hyperspectral camera by considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin.

[0050] The function and / or operation of the oxygen saturation prediction program (200) will be examined in detail with reference to FIG. 2.

[0051] FIG. 2 is a block diagram conceptually illustrating the function of an oxygen saturation prediction program according to an embodiment of the present invention.

[0052] Referring to FIG. 2, the oxygen saturation prediction program (200) may include a hyperspectral image acquisition unit (210), an absorbance measurement unit (220), a wavelength band determination unit (230), and an oxygen saturation prediction unit (240).

[0053] The hyperspectral image acquisition unit (210), absorbance measurement unit (220), wavelength band determination unit (230), and oxygen saturation prediction unit (240) illustrated in FIG. 2 conceptually divide the functions of the oxygen saturation prediction program (200) to easily explain the functions of the oxygen saturation prediction program (200), but are not limited thereto. According to embodiments, the functions of the hyperspectral image acquisition unit (210), absorbance measurement unit (220), wavelength band determination unit (230), and oxygen saturation prediction unit (240) can be merged / separated, and can also be implemented as a series of commands included in one program.

[0054] First, the hyperspectral image acquisition unit (210) can acquire a hyperspectral image of the subject using a hyperspectral camera.

[0055] Specifically, the hyperspectral image acquisition unit (210) can acquire a hyperspectral image containing only a diffuse reflection component by using a first polarizing filter attached to a hyperspectral camera and a second polarizing filter attached to a light source (e.g., an LED light source) so as to be orthogonal to the direction of the first polarizing filter.

[0056] Accordingly, by removing the surface reflection component that is sensitive to external environmental factors, it becomes possible to accurately reflect information on the skin dermis layer, which is closely related to oxygen saturation.

[0057] Next, the absorbance measuring unit (220) can measure absorbance based on a wavelength band that includes only diffuse reflection components from a hyperspectral image, and for example, the absorbance can be expressed as in the following mathematical expression 1.

[0058]

[0059] Here, can mean absorbance, may mean a diffuse reflection component.

[0060] Meanwhile, the absorbance measurement unit (220) can measure the absorbance of at least one facial area detected from the hyperspectral image by considering the correlation with the change in oxygen saturation in the facial area of ​​the subject.

[0061] Here, the absorbance measurement unit (220) can detect at least one facial area by considering the optical properties of oxygen saturation based on an image processing algorithm.

[0062] For example, the at least one facial region may include a forehead region, an under-eye region, a cheek region, and an under-nose region, which are regions sensitive to changes in oxygen saturation.

[0063] In addition, at least one face region can be set as a region of interest by the user, and the absorbance measurement unit (220) can detect at least one face region corresponding to the region of interest from a hyperspectral image using yolov5.

[0064] Meanwhile, the absorbance measurement unit (220) can remove shade elements of the face area of ​​the subject by considering the geometric structure of the face area of ​​the subject.

[0065] For example, the absorbance measurement unit (220) can remove the shadow element of the face area of ​​the subject by removing the difference value between the three-dimensional plane generated through MCML simulation and the absorbance measured for at least one face area.

[0066] This allows for more precise measurement of oxygen saturation by eliminating differences in absorbance that occur depending on the different facial shapes of each person.

[0067] Meanwhile, removing the shadow elements of the subject's face area through MCML simulation will be described in detail later in Fig. 5.

[0068] Next, the wavelength band determination unit (230) can determine at least two wavelength bands among the wavelength bands measured using a hyperspectral camera by considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin. Here, the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin can be measured in the skin dermis layer.

[0069] Specifically, the wavelength band determination unit (230) can determine at least two wavelength bands that are sensitive to changes in oxygen saturation by considering the rate of change in absorbance for oxyhemoglobin and deoxyhemoglobin measured for each wavelength band in a hyperspectral camera.

[0070] More specifically, the wavelength band determination unit (230) can determine at least two wavelength bands based on a first condition that the difference between the absorbances of deoxyhemoglobin for at least two wavelength bands is less than a first threshold value approaching 0, and a second condition that the absorbances of oxyhemoglobin and deoxyhemoglobin each exceed a second threshold value approaching 0. Here, the first threshold value and the second threshold value can be preset by a user.

[0071] Meanwhile, the wavelength band determination unit (230) can remove the blood flow component based on the ratio of absorbance measured in at least two determined wavelength bands.

[0072] Specifically, the absorbance of the skin dermis layer in the measurement wavelength band is determined based on the product of the absorption coefficient in the skin dermis layer and the measurement wavelength band, and can be expressed as in the following mathematical expression 2.

[0073]

[0074] Here, can mean the absorption coefficient of the skin dermis layer, can mean the absorbance of the skin dermis layer in the measurement wavelength band, may refer to blood flow components, can mean oxygen saturation, may refer to the absorption coefficient of oxyhemoglobin, may refer to the absorbance of oxyhemoglobin in the measurement wavelength band, may refer to the absorption coefficient of deoxyhemoglobin, may refer to the absorbance of deoxyhemoglobin in the measurement wavelength band.

[0075] In addition, when the at least two wavelength bands determined above are assumed to be a first measurement wavelength band and a second measurement wavelength band, the wavelength band determination unit (230) can determine a ratio between the first absorbance of the skin dermis layer in the first measurement wavelength band and the second absorbance of the skin dermis layer in the second measurement wavelength band, and can be expressed as in the following mathematical expression 3.

[0076]

[0077] Here, may mean the first absorbance, may mean the second absorbance, may mean the ratio of the first absorbance and the second absorbance.

[0078] In this way, the wavelength band determination unit (230) can remove the blood flow component by using the ratio of absorbance measured in at least two wavelength bands, and by removing the blood flow component that affects the diffuse reflection component, the effect of enabling more precise prediction of oxygen saturation can be achieved.

[0079] Next, the oxygen saturation prediction unit (240) can predict the oxygen saturation of the subject based on the absorbance for at least one facial area and at least two wavelength bands.

[0080] Specifically, the oxygen saturation prediction unit (240) can predict the oxygen saturation for each of at least one facial region based on the absorbance measured time-seriesly for each of at least one facial region and at least two determined wavelength bands. Here, the oxygen saturation can be expressed as in the following mathematical expression 4, and mathematical expression 4 can mean mathematical expression 3 organized based on oxygen saturation.

[0081]

[0082] More specifically, the oxygen saturation prediction unit (240) can predict the oxygen saturation for each of at least one facial region corresponding to the region of interest from absorbance measured in time series (e.g., absorbance measured every 4 seconds) and at least two wavelength bands using a first deep learning model (e.g., Auto Encoder LSTM (long short term memory)).

[0083] Additionally, the oxygen saturation prediction unit (240) can predict the oxygen saturation of the subject by using the oxygen saturation for each of at least one facial areas corresponding to the area of ​​interest.

[0084] Specifically, the oxygen saturation prediction unit (240) can predict the oxygen saturation of the subject from the oxygen saturation of each of at least one facial region corresponding to the region of interest using a second deep learning model (e.g., a fusion deep learning network).

[0085] Figure 3 is a flowchart illustrating a method for predicting oxygen saturation according to one embodiment of the present invention.

[0086] Referring to FIGS. 2 and 3, the hyperspectral image acquisition unit (210) can acquire a hyperspectral image of a subject using a hyperspectral camera (S310).

[0087] Next, the absorbance measuring unit (220) can measure the absorbance of at least one facial area detected by considering the correlation with the change in oxygen saturation in the facial area of ​​the subject from the hyperspectral image (S320).

[0088] Next, the wavelength band determination unit (230) can determine at least two wavelength bands among the wavelength bands measured using the hyperspectral camera by considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin (S330).

[0089] Additionally, the oxygen saturation prediction unit (240) can predict the oxygen saturation of the subject based on the absorbance for at least one facial area and at least two wavelength bands (S340).

[0090] FIG. 4 is a drawing exemplarily showing obtaining only a diffuse reflection component using a polarizing filter attached to a hyperspectral camera and lighting according to one embodiment of the present invention.

[0091] Referring to FIGS. 2 and 4, the hyperspectral image acquisition unit (210) can acquire a hyperspectral image of a subject using a hyperspectral camera.

[0092] Figure 4 illustrates the epidermis and dermis of the skin, and the surface reflective component includes the absorbance of melanin, which is sensitive to changes in external environmental factors, thus hindering accurate prediction of oxygen saturation.

[0093] Accordingly, the hyperspectral image acquisition unit (210) can acquire a hyperspectral image that includes only the diffuse reflection component of the skin dermis layer by using the first polarizing filter (401) attached to the hyperspectral camera and the second polarizing filter (402) attached to the LED light so as to be orthogonal to the direction of the first polarizing filter (401).

[0094] Here, the first polarizing filter (401) and the second polarizing filter (402) can be installed at one end of the hyperspectral camera and the LED light, respectively.

[0095] FIG. 5 is a drawing exemplarily showing the removal of shadow elements from a face area of ​​a subject according to one embodiment of the present invention.

[0096] Referring to FIGS. 2 and 5, the absorbance measurement unit (220) can remove shadow elements from the face area of ​​the subject by considering the geometric structure of the face area of ​​the subject.

[0097] Specifically, the absorbance measurement unit (220) can remove the shadow element (503) of the face area of ​​the subject by removing the difference value between the three-dimensional plane (502) generated through MCML simulation and the absorbance (501) measured for at least one face area. Accordingly, the absorbance can be accurately measured regardless of the shape of the person's face, and can be expressed as in the following mathematical expression 5.

[0098]

[0099] Here, can mean the absorbance with the shadow elements of the face area of ​​the subject removed, can mean absorbance, can mean the difference value between the three-dimensional plane generated through MCML simulation and the absorbance measured for the face area of ​​the subject.

[0100] In this regard, a three-dimensional plane (502) for removing shadow elements (503) of the face area of ​​the subject can be determined using an MCML simulator.

[0101] Here, since the extinction coefficient is affected by melanin, blood flow, and oxygen saturation, the extinction coefficients of melanin, oxyhemoglobin, and deoxyhemoglobin can be modeled through MCML simulation based on the settings for the volume of melanin in the epidermal layer of the skin, the volume of hemoglobin in the dermal layer of the skin, the blood flow, and the skin thickness.

[0102] And, a three-dimensional plane (502) can be determined based on the absorption coefficients of the melanin, oxyhemoglobin, and deoxyhemoglobin and at least two wavelength bands (e.g., 409.8 μm, 574.9 μm, and 710.1 μm).

[0103] More specifically, the absorbance measurement unit (220) can determine the distance between the (1, 1, 1) vector (501) passing through the measured absorbance and the three-dimensional plane (502) generated through MCML simulation as a shading element (503).

[0104] In addition, the absorbance measuring unit (220) can measure absorbance regardless of the shape of a person's face by removing the value corresponding to the shaded element (503) from the measured absorbance (501).

[0105] This allows for more precise measurement of oxygen saturation by eliminating differences in absorbance that occur depending on the different facial shapes of each person.

[0106] FIG. 6 is a drawing exemplarily showing determining at least two wavelength bands for predicting oxygen saturation by considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin according to one embodiment of the present invention.

[0107] Referring to FIGS. 2 and 6, the wavelength band determination unit (230) can determine at least two wavelength bands among the wavelength bands measured using a hyperspectral camera by considering the absorbance (602) of oxyhemoglobin and the absorbance (601) of deoxyhemoglobin.

[0108] Specifically, the wavelength band determination unit (230) can determine at least two wavelength bands that are sensitive to changes in oxygen saturation by considering the rate of change in absorbance (602) of oxyhemoglobin and absorbance (601) of deoxyhemoglobin measured for each wavelength band in a hyperspectral camera.

[0109] More specifically, the wavelength band determining unit (230) can determine at least two wavelength bands that satisfy a first condition that the difference between the absorbances (601) of deoxyhemoglobin for at least two wavelength bands is less than a first threshold value approaching 0.

[0110] Additionally, the wavelength band determining unit (230) can determine at least two wavelength bands that satisfy a second condition regarding that the absorbance of oxyhemoglobin (602) and the absorbance of deoxyhemoglobin (601) each exceed a second threshold value approaching 0.

[0111] That is, the wavelength band determining unit (230) can determine at least two wavelength bands (e.g., 528 μm, 554 μm, 565 μm) that satisfy the first condition and the second condition by considering the change rate of the absorbance (602) of oxyhemoglobin and the absorbance (601) of deoxyhemoglobin.

[0112] The combination of each block of the block diagram and each step of the flowchart attached to the present invention may be performed by computer program instructions. These computer program instructions may be installed in an encoding processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, so that the instructions executed by the encoding processor of the computer or other programmable data processing equipment create a means for performing the functions described in each block of the block diagram or each step of the flowchart. These computer program instructions may also be stored in a computer-available or computer-readable memory that can direct a computer or other programmable data processing equipment to implement the functions in a specific manner, so that the instructions stored in the computer-available or computer-readable memory can also produce an article of manufacture that includes an instruction means for performing the functions described in each block of the block diagram or each step of the flowchart. Since the computer program instructions can also be installed on a computer or other programmable data processing device, a series of operational steps are performed on the computer or other programmable data processing device to create a computer-executable process, and the instructions that cause the computer or other programmable data processing device to perform the steps for executing the functions described in each block of the block diagram and each step of the flowchart can also provide steps for executing the functions described in each block of the block diagram and each step of the flowchart.

[0113] Additionally, each block or step may represent a module, segment, or portion of code that includes one or more executable instructions for performing a specific logical function(s). It should also be noted that in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps depicted in succession may actually be performed substantially concurrently, or the blocks or steps may sometimes be performed in reverse order, depending on the functionality they perform.

[0114] The above description is merely an illustrative illustration of the technical idea of ​​the present invention, and those skilled in the art will appreciate that various modifications and variations can be made without departing from the essential quality of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate, rather than limit, the technical idea of ​​the present invention, and the scope of the technical idea of ​​the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of the rights of the present invention.

Claims

1. A non-contact oxygen saturation prediction method based on a hyperspectral camera, A step of obtaining a hyperspectral image of a subject using the hyperspectral camera; A step of measuring the absorbance of at least one facial region detected from the hyperspectral image by considering the correlation with the change in the oxygen saturation among the facial regions of the subject; A step of determining at least two optimal wavelength bands among the wavelength bands measured using the hyperspectral camera, taking into account the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and A step of predicting the oxygen saturation of the subject based on the absorbance for at least one facial area and the at least two optimal wavelength bands. Method for predicting oxygen saturation.

2. In paragraph 1, The step of acquiring the above hyperspectral image is: A step of obtaining a hyperspectral image including only a diffuse reflection component by using a first polarizing filter attached to the hyperspectral camera and a second polarizing filter attached to the light so as to be orthogonal to the direction of the first polarizing filter. Method for predicting oxygen saturation.

3. In paragraph 1, The step of measuring the above absorbance is: Considering the geometric structure of the face area of ​​the subject, a step of removing shadow elements of the face area of ​​the subject is included. Method for predicting oxygen saturation.

4. In paragraph 1, In the step of measuring the absorbance, At least one facial region detected in consideration of the correlation with the change in the oxygen saturation is detected based on an image processing algorithm, The at least one facial region includes at least one of a forehead region, an under-eye region, a cheek region, and an under-nose region. Method for predicting oxygen saturation.

5. In paragraph 1, The step of predicting the above oxygen saturation is: A step of removing a blood flow component based on the ratio of absorbance measured in at least two optimal wavelength bands determined above; and Further comprising a step of predicting the oxygen saturation based on at least two optimal wavelength bands from which the blood flow component has been removed. Method for predicting oxygen saturation.

6. In paragraph 1, The step of predicting the above oxygen saturation is: A step of predicting oxygen saturation for each of the at least one facial region based on absorbance measured time-seriesly for each of the at least one facial region and the at least two optimal wavelength bands; and A step of predicting oxygen saturation for the subject using oxygen saturation for each of the at least one facial region. Method for predicting oxygen saturation.

7. A non-contact oxygen saturation prediction device based on a hyperspectral camera, Memory in which the oxygen saturation prediction program is stored; and A processor that loads the oxygen saturation prediction program from the memory and executes the oxygen saturation prediction program, The above processor, Obtaining a hyperspectral image of the subject using the above hyperspectral camera, Measure the absorbance of at least one facial region detected from the hyperspectral image by considering the correlation with the change in oxygen saturation among the facial regions of the subject, Considering the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin, at least two optimal wavelength bands among the wavelength bands measured using the hyperspectral camera are determined, Predicting the oxygen saturation of the subject based on the absorbance for at least one facial area and the at least two optimal wavelength bands. Oxygen saturation prediction device.

8. In paragraph 7, The above processor, A hyperspectral image containing only diffuse reflection components is obtained by using a first polarizing filter attached to the hyperspectral camera and a second polarizing filter attached to the light so as to be orthogonal to the direction of the first polarizing filter. Oxygen saturation prediction device.

9. In paragraph 7, The above processor, Considering the geometric structure of the face area of ​​the subject, the shadow elements of the face area of ​​the subject are removed. Oxygen saturation prediction device.

10. In paragraph 7, At least one facial region detected in consideration of the correlation with the change in the oxygen saturation is detected based on an image processing algorithm, The at least one facial region includes at least one of a forehead region, an under-eye region, a cheek region, and an under-nose region. Oxygen saturation prediction device.

11. In paragraph 7, The above processor, The blood flow component is removed based on the ratio of the absorbance measured in the at least two optimal wavelength bands determined above, and the oxygen saturation is predicted based on the at least two optimal wavelength bands from which the blood flow component is removed. Oxygen saturation prediction device.

12. In paragraph 7, The above processor, Predicting oxygen saturation for each of said at least one facial region based on absorbance measured time-seriesly for each of said at least one facial region and said at least two optimal wavelength bands, Predicting oxygen saturation for the subject using oxygen saturation for each of the at least one facial region Oxygen saturation prediction device.

13. A non-transitory computer-readable recording medium storing a computer program, The above computer program, when executed by a processor, A step of acquiring a hyperspectral image of a subject using a hyperspectral camera; A step of measuring the absorbance of at least one facial region detected from the hyperspectral image by considering the correlation with the change in the oxygen saturation among the facial regions of the subject; A step of determining at least two optimal wavelength bands among the wavelength bands measured using the hyperspectral camera, taking into account the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and A step of predicting the oxygen saturation of the subject based on the absorbance for at least one facial area and the at least two optimal wavelength bands. A non-contact oxygen saturation prediction method based on a hyperspectral camera is provided, comprising instructions for causing the processor to perform the method. Non-transitory computer-readable recording medium.

14. A computer program stored in a non-transitory computer-readable recording medium, The above computer program, when executed by a processor, A step of acquiring a hyperspectral image of a subject using a hyperspectral camera; A step of measuring the absorbance of at least one facial region detected from the hyperspectral image by considering the correlation with the change in the oxygen saturation among the facial regions of the subject; A step of determining at least two optimal wavelength bands among the wavelength bands measured using the hyperspectral camera, taking into account the absorbance of oxyhemoglobin and the absorbance of deoxyhemoglobin; and A step of predicting the oxygen saturation of the subject based on the absorbance for at least one facial area and the at least two optimal wavelength bands. A non-contact oxygen saturation prediction method based on a hyperspectral camera is provided, comprising instructions for causing the processor to perform the method. Computer program.

Images