JP2023028182A - Measuring device, measuring method, and program - Google Patents

Abstract

[Problem] To provide a measuring device capable of estimating blood vessel type in a non-contact manner. [Solution] The measuring device includes an image acquisition unit that acquires at least two images including blood vessel images of an eyeball captured by irradiating at least two types of light having different peak wavelengths, a blood vessel cross-sectional profile acquisition unit that acquires a plurality of blood vessel cross-sectional profiles each showing a distribution of pixel values of pixels at a plurality of blood vessel cross-sectional positions for each of the at least two images, an absorbance calculation unit that calculates the absorbance at each of the plurality of blood vessel cross-sectional positions from the plurality of blood vessel cross-sectional profiles, a distribution creation unit that creates a plurality of two-dimensional distributions, each set including an absorbance ratio and a blood vessel contrast index, and the absorbance ratio is calculated from the absorbance at each blood vessel cross-sectional position, and a blood vessel type estimation unit that clusters the two-dimensional distribution to obtain clusters of a plurality of blood vessel types and estimates the blood vessel types at the plurality of blood vessel cross-sectional positions. [Selected Figure] Figure 2

Description

The present disclosure relates to a measuring device, measuring method and program.

Non-Patent Document 1 discloses a technique for identifying blood vessels from spectroscopic images obtained using a halogen lamp and an image replicating imaging spectrometer (IRIS) and calculating the correlation with oxygen saturation. .

MacKenzie, Lewis Edward (2016) In vivo microvascular oximetry using multispectral imaging. PhD thesis. (http://theses.gla.ac.uk/7732/)

For example, because blood flowing through the epiconjunctival artery of the eye exchanges oxygen with air, the epiconjunctival artery is always relatively saturated with oxygen. On the other hand, little oxygen exchange takes place between blood and air flowing through the episcleral artery. Therefore, the oxygen saturation of the episcleral artery is a reflection of a person's state of health. In other words, the oxygen saturation differs depending on the type of blood vessel. Therefore, in order to estimate the oxygen saturation reflecting a person's health condition in a non-contact manner, it is desired to identify the blood vessel type in a non-contact manner. With the technique disclosed in Non-Patent Document 1, since it is not possible to distinguish between types of blood vessels such as arteries and veins, it may not be possible to estimate the oxygen saturation for each type of blood vessel. Accordingly, an object of one aspect of the present disclosure is to provide a measuring device, a measuring method, and a program that can estimate the blood vessel type without contact.

A measurement device according to an aspect of the present disclosure includes an image acquisition unit that acquires at least two images including a blood vessel image of an eyeball captured by irradiating at least two types of light having different peak wavelengths; a blood vessel cross-section profile obtaining unit for obtaining a plurality of blood vessel cross-section profiles respectively indicating distributions of pixel values of pixels at a plurality of blood vessel cross-section positions for each of the two images; and a plurality of blood vessel contrast indexes each including an absorbance ratio of each blood vessel cross-section position and a blood vessel contrast index indicating the clarity of each blood vessel cross-section position of the blood vessel image. a set of two-dimensional distributions, wherein the absorbance ratio is calculated from the absorbance at each of the blood vessel cross-sectional positions for the at least two images; a blood vessel type estimating unit that obtains type clusters and estimates the blood vessel types at the plurality of blood vessel cross-sectional positions.

A measurement method according to an aspect of the present disclosure includes a step of acquiring at least two images including a blood vessel image of an eyeball captured by irradiating at least two types of light having peak wavelengths different from each other; acquiring a plurality of blood vessel cross-section profiles each showing a distribution of pixel values of pixels at a plurality of blood vessel cross-section positions for each of the images; calculating the absorbance of the locations; and creating a plurality of sets of two-dimensional distributions, each set including an absorbance ratio of each blood vessel cross-section position and a blood vessel contrast index indicating the clarity of each blood vessel cross-section position of the blood vessel image. and calculating the absorbance ratio from the absorbance at each of the blood vessel cross-sectional positions for the at least two images; and estimating the type of blood vessel at the cross-sectional position.

A program according to an aspect of the present disclosure includes a function of acquiring at least two images including a blood vessel image of an eyeball captured by irradiating a computer with at least two types of light having mutually different peak wavelengths; a function of obtaining a plurality of blood vessel cross-sectional profiles respectively indicating distributions of pixel values of pixels at a plurality of blood vessel cross-sectional positions for each of the two images; A plurality of sets of two-dimensional distributions each including a function of calculating the absorbance of the blood vessel cross-section position, and a blood vessel contrast index, each set of which indicates the absorbance ratio of the blood vessel cross-section position and the clarity of the blood vessel cross-section position of the blood vessel image. and the absorbance ratio is a function calculated from the absorbance at each of the blood vessel cross-sectional positions for the at least two images, and the two-dimensional distribution is clustered to obtain clusters of a plurality of types of blood vessels, and the plurality of and a function of estimating the type of blood vessel at the cross-sectional position of the blood vessel.

It is a figure which shows an example of a measurement system. It is a figure which shows an example of a structure of a measuring device. It is a flow chart which shows an example of processing performed with a measuring device concerning a first embodiment. It is a figure which shows an example of the image imaged by irradiating an eyeball with two types of light which have a different peak wavelength. FIG. 4 is a diagram for explaining processing for determining a blood vessel cross-section position; FIG. 4 is a diagram showing an example of a plurality of blood vessel cross-sectional positions included in a blood vessel region; FIG. 4 is a diagram showing an example of a blood vessel cross-sectional profile; 4 is a flowchart showing an example of processing executed by the measuring device according to the first embodiment following FIG. 3; It is a figure which shows an example of a two-dimensional distribution. It is a flow chart which shows an example of processing performed with a measuring device concerning a second embodiment. It is a flow chart which shows an example of processing performed with a measuring device concerning a third embodiment. It is a flow chart which shows an example of processing performed with a measuring device concerning a fourth embodiment.

(First embodiment)
A first embodiment will be described with reference to FIGS. 1 to 9. FIG. In the drawings, the same or similar elements are denoted by the same reference numerals, and overlapping descriptions are omitted.

FIG. 1 is a diagram showing an example of a measurement system 100. As shown in FIG. The measurement system 100 includes a measurement device 101, a light source 102a, and a light source 102b.

The measuring device 101 captures an image of the eyeball 103 in an environment where the eyeball 103 is irradiated with light emitted from the light source 102a and the light source 102b, and estimates the oxygen saturation of the blood vessels of the eyeball 103 from the captured image of the eyeball 103. . Oxygen saturation is an index indicating the ratio of the amount of hemoglobin bound to oxygen to the amount of hemoglobin contained in red blood cells contained in arterial blood. Oxygen saturation decreases when the ability of the body to take in oxygen decreases due to diseases such as the lungs and heart.

The light source 102a is arranged so that the emitted light 104a emitted from the light source 102a is irradiated to the eyeball 103 of the living body. Similarly, the light source 102b is arranged so that the emitted light 104b emitted from the light source 102b irradiates the eyeball 103 of the living body. The light sources 102a and 102b are configured by LEDs (Light Emitting Diodes) that emit emitted light 104a and emitted light 104b having peak wavelengths different from each other.

For example, the light source 102a and the light source 102b emit emitted light 104a and emitted light 104b having peak wavelengths at which there is a difference between the molar extinction coefficient of oxygenated hemoglobin and the molar extinction coefficient of reduced hemoglobin. Alternatively, light source 102a emits emitted light 104a having a peak wavelength at which there is a difference between the molar extinction coefficient of oxygenated hemoglobin and the molar absorption coefficient of reduced hemoglobin, and the light source 102b emits emitted light 104a having a difference between the molar extinction coefficient of oxygenated hemoglobin and the molar absorption coefficient of reduced hemoglobin. The emitted light 104b having a peak wavelength that does not cause any difference from the coefficient may be emitted.

For example, the light sources 102a and 102b emit emitted light 104a and emitted light 104b at different timings. As a result, emitted light 104a and emitted light 104b emitted from light source 102a and light source 102b enter eyeball 103 at different timings. The emitted light 104 a and the emitted light 104 b are reflected by the eyeball 103 . Thereby, reflected light 105a and reflected light 105b of emitted light 104a and emitted light 104b are generated.

FIG. 2 is a diagram showing an example of the configuration of the measuring device 101. As shown in FIG. The measurement device 101 includes an imaging unit 201, a storage unit 202, a control unit 203, a processing unit 204, and the like.

The imaging unit 201 receives the reflected light 105 a and the reflected light 105 b and captures an image of the eyeball 103 . For example, the imaging unit 201 is configured by a CCD (Charged Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor) image sensor. For example, the imaging unit 201 may be configured by a camera image sensor including RGB (Red Green Blue) filters.

The storage unit 202 is a recording medium capable of recording various data, programs, etc., and is configured by, for example, a hard disk, an SSD (Solid State Drive), a semiconductor memory, or the like.

The control unit 203 controls the light emission timings of the light sources 102a and 102b. Furthermore, the control unit 203 controls the shutter operation of the imaging unit 201 . The control unit 203 may control the opening time of the shutter of the imaging unit 201 to cause the imaging unit 201 to image the eyeball 103 at a predetermined frame rate for a predetermined time.

The processing unit 204 executes various processes according to programs and data stored in the storage unit 202 . The processing unit 204 is realized by, for example, a processor such as a CPU (Central Processing Unit).

The processing unit 204 includes an image acquisition unit 205, a blood vessel region extraction unit 206, a blood vessel cross section determination unit 207, a blood vessel cross section profile acquisition unit 208, a suitable distribution identification unit 209, an absorbance calculation unit 210, an absorbance ratio calculation unit 211, and a blood vessel contrast index calculation. It includes a unit 212, a distribution creating unit 213, a blood vessel type estimating unit 214, an oxygen saturation calculating unit 215, and the like.

The image acquisition unit 205 acquires at least two images 221a and 221b including blood vessel images of the eyeball captured by irradiating at least two types of light having different peak wavelengths.

The blood vessel region extraction unit 206 extracts a blood vessel region 222a by performing image processing on the image 221a. A blood vessel region 222a includes a blood vessel image. Similarly, the blood vessel region extraction unit 206 extracts a blood vessel region 222b by performing image processing on the image 221b. The blood vessel region 222b includes a blood vessel image similar to the blood vessel region 222a.

The blood vessel cross-section determination unit 207 determines a plurality of blood vessel cross-section positions from the blood vessel region 222a. Similarly, the blood vessel cross-section determination unit 207 determines a plurality of blood vessel cross-section positions from the blood vessel region 222b.

The blood vessel cross-sectional profile acquisition unit 208 acquires multiple blood vessel cross-sectional profiles 223a and multiple blood vessel cross-sectional profiles 223b for each of at least two images 221a and 221b. A plurality of blood vessel cross-section profiles 223a respectively indicate pixel value distributions of pixels at a plurality of blood vessel cross-section positions in the blood vessel region 222a. A plurality of blood vessel cross-section profiles 223b respectively indicate pixel value distributions of pixels at a plurality of blood vessel cross-section positions in the blood vessel region 222b.

The matching distribution specifying unit 209 specifies a matching distribution 224a that matches the blood vessel cross-sectional profile 223a. For example, the matching distribution identifying unit 209 identifies a matching distribution 224a that exhibits a normal distribution by performing Gaussian fitting on the blood vessel cross-sectional profile 223a. Similarly, the matching distribution identification unit 209 identifies a matching distribution 224b that matches the blood vessel cross-sectional profile 223b.

The absorbance calculator 210 calculates the absorbance 225a of each blood vessel cross-sectional position among the multiple blood vessel cross-sectional positions acquired for the image 221a from the multiple blood vessel cross-sectional profiles 223a. Similarly, the absorbance calculator 210 calculates the absorbance 225b of each blood vessel cross-sectional position among the multiple blood vessel cross-sectional positions acquired for the image 221b from the multiple blood vessel cross-sectional profiles 223b.

The absorbance ratio calculator 211 calculates the absorbance ratio 226 from the absorbance 225a and the absorbance 225b of each blood vessel cross-sectional position calculated for at least two images 221a and 221b.

The blood vessel contrast index calculation unit 212 indicates the clarity of each blood vessel cross-sectional position of the blood vessel image from the blood vessel cross-sectional profile 223a or the blood vessel cross-sectional profile 223b obtained for at least one of the at least two images 221a and 221b. A blood vessel contrast index 227 is calculated.

The distribution creation unit 213 creates a plurality of sets of two-dimensional distributions 228, each set including the absorbance ratio 226 of each blood vessel cross-section position and the blood vessel contrast index 227 of each blood vessel cross-section position.

The blood vessel type estimation unit 214 clusters the two-dimensional distribution 228 to obtain clusters of a plurality of blood vessel types, and estimates the blood vessel types at a plurality of blood vessel cross-sectional positions.

The oxygen saturation calculator 215 calculates the oxygen saturation from the absorbance ratios 226 included in the groups belonging to a specific cluster among the plurality of types of clusters.

FIG. 3 is a flowchart showing an example of processing executed by the measuring device 101. As shown in FIG. When the measurement device 101 is activated, the control unit 203 enters a state in which it can start controlling the shutter operation of the imaging unit 201 . In this example, it is assumed that the emitted light 104a is light having a peak wavelength _λ1 . Further, the peak wavelength λ ₁ is assumed to be a wavelength at which a difference occurs between the molar extinction coefficients of oxygenated hemoglobin and reduced hemoglobin. It is also assumed that the emitted light 104b is light having a peak wavelength λ ₂ =570 nm. A peak wavelength of 570 nm is a wavelength at which there is no difference between the molar extinction coefficients of oxygenated hemoglobin and reduced hemoglobin.

In step S301, the image acquisition unit 205 acquires images 221a and 221b captured by irradiating the eyeball 103 with at least two types of light having at least two types of peak wavelengths. For example, when the imaging unit 201 captures an image of the eyeball 103 by receiving the reflected light 105a at a predetermined frame rate for a predetermined time, the image acquisition unit 205 includes the captured image of the eyeball 103. Acquire moving images. In that case, the image acquiring unit 205 extracts the image 221a from the acquired moving image. Similarly, the image acquiring unit 205 extracts the image 221b from the moving image including the image of the eyeball 103 captured by receiving the reflected light 105b.

Image 221 a and image 221 b include images of the same plurality of blood vessels of eyeball 103 . The eyeball 103 has a conjunctiva and a sclera. The conjunctiva is the part of the mucous membrane that covers the back of the eyelid and the surface of the eyeball 103 to the periphery of the iris. The sclera is the underlying layer of the conjunctiva, makes up the outer membrane of the eye, and is clear and colorless. The sclera makes up the white of the eye. Therefore, when the white region of the eyeball 103 is imaged, the image 221a and the image 221b include an image of the suprascleral blood vessels and an image of the supraconjunctival blood vessels.

For example, when the imaging unit 201 is configured by an image sensor for a camera including RGB (Red Green Blue) filters, the images 221a and 221b are composed of only green G pixels among RGB pixels. It may be an image showing a value. In addition, the image 221a and the image 221b may represent RAW data of pixels of G, which is green. RAW data is original luminance data that has not undergone compression processing.

In step S302, the blood vessel region extraction unit 206 extracts the blood vessel region 222a from the image 221a. For example, the blood vessel region extracting unit 206 performs image processing such as edge detection processing on the image 221a to identify the blood vessel image. The blood vessel region extraction unit 206 then identifies a region including the identified blood vessel image from the image 221a as a blood vessel region 222a. Similarly, in step S302, the blood vessel region extraction unit 206 extracts the blood vessel region 222b from the image 221b.

FIG. 4 is a diagram showing an example of an image captured by irradiating the eyeball 103 with two types of light having different peak wavelengths. The image 221a is an image captured by irradiating the eyeball 103 with the emitted light 104a. The image 221b is an image captured by irradiating the eyeball 103 with the emitted light 104b. A blood vessel region 222a includes a blood vessel image included in the image 221a. Also, the blood vessel region 222b includes a blood vessel image at the same position as the blood vessel region 222a included in the image 221a.

In subsequent step S303, the blood vessel cross-section determination unit 207 determines a plurality of blood vessel cross-section positions from the blood vessel region 222a. For example, the blood vessel cross-section determination unit 207 performs principal component analysis on the coordinate values of the pixels of the outline of the blood vessel image in the blood vessel region 222a, and determines the direction of the largest variance, which is the first principal component, of the blood vessel image. It is specified as the blood vessel direction in which the blood vessel to which it belongs extends. Then, the blood vessel cross-section determination unit 207 determines a position along the normal direction that intersects the identified blood vessel direction as the blood vessel cross-section position. For example, when the size of the blood vessel region 222a is 300 pixels×300 pixels, the blood vessel cross-section determination unit 207 determines 1000 blood vessel cross-section positions from the blood vessel region 222a. Similarly, in step S303, the blood vessel cross-section determination unit 207 determines a plurality of blood vessel cross-section positions from the blood vessel region 222b.

FIG. 5 is a diagram for explaining the process of determining the blood vessel cross-sectional position. A region 501 illustrated in FIG. 5 indicates a blood vessel image. An arrow 502 indicates the direction of the blood vessel to which the blood vessel image indicated by the area 501 belongs. A line segment 503 indicates the blood vessel cross-sectional position. The blood vessel cross-sectional position is a range in a direction that intersects the blood vessel direction, and is determined across blood vessel images. For example, the blood vessel cross-sectional position is the range in the direction orthogonal to the blood vessel direction.

FIG. 6 is a diagram showing an example of a plurality of blood vessel cross-sectional positions 601 included in the blood vessel region 222a.

In subsequent step S304, the blood vessel cross-sectional profile acquisition unit 208 acquires the blood vessel cross-sectional profile 223a for each of the plurality of blood vessel cross-sectional positions in the image 221a. Similarly, in step S304, the blood vessel cross-sectional profile acquisition unit 208 acquires the blood vessel cross-sectional profile 223b for each of the multiple blood vessel cross-sectional positions in the image 221b.

In step S305, the distribution generating unit 213 selects the target blood vessel cross-section position included in the image 221a from a plurality of blood vessel cross-section positions of the image 221a. Similarly, in step S305, the distribution generator 213 selects target blood vessel cross-section positions included in the image 221b from a plurality of blood vessel cross-section positions in the image 221b. The blood vessel image to which the target blood vessel cross-section position selected from the image 221b belongs is the blood vessel image at the same position as the target blood vessel cross-section position selected from the image 221a.

In step S306, the matching distribution identifying unit 209 identifies the matching distribution 224a that matches the blood vessel cross-sectional profile 223a of the target blood vessel cross-sectional position included in the image 221a. Specifically, the matching distribution specifying unit 209 specifies a matching distribution 224a that matches the pixel value distribution indicated by the blood vessel cross-sectional profile 223a at the target blood vessel cross-sectional position included in the image 221a. For example, the matching distribution identifying unit 209 identifies a matching distribution 224a that exhibits a normal distribution by performing Gaussian fitting on the blood vessel cross-sectional profile 223a. Similarly, in step S306, the matching distribution specifying unit 209 specifies a matching distribution 224b that matches the blood vessel cross-sectional profile 223b of the target blood vessel cross-sectional position included in the image 221b.

FIG. 7 is a diagram showing an example of a blood vessel cross-sectional profile 223a. In FIG. 7, the horizontal axis is the position and the vertical axis is the pixel value. Each plotted point 701 indicates the pixel value of each pixel on the blood vessel cross-section position included in the blood vessel cross-section profile 223a. Iv indicates the minimum pixel value of the matching distribution 224a. Io is the baseline of the fitted distribution 224a. |A| is the absolute value of the difference between the pixel value Iv and the pixel value on the baseline Io at the position of the pixel value Iv. Therefore, |A| indicates the height of the peak of the pixel value in the matching distribution 224a. σ is a value indicating the variation from the average value of the pixel values indicated by the blood vessel cross-sectional profile 223a. 2σ illustrated in FIG. 7 indicates the blood vessel diameter at the target blood vessel cross-sectional position of the blood vessel cross-sectional profile 223a. Therefore, |A| indicates the difference between the brightness of the inner center of the blood vessel wall and the brightness of the outer side of the blood vessel wall.

In subsequent step S307, the absorbance calculator 210 calculates the absorbance 225a at the target blood vessel cross-section position based on the blood vessel cross-section profile 223a at the target blood vessel cross-section position included in the image 221a. Similarly, in step S307, the absorbance calculator 210 calculates the absorbance 225b at the target blood vessel cross-section position based on the blood vessel cross-section profile 223b at the target blood vessel cross-section position included in the image 221b.

OD shown in Equation (1) is the absorbance at the target blood vessel cross-sectional position included in the image 221a or image 221b. β is the pixel value on the baseline Io at the position of the minimum pixel value Iv in the matching distribution 224a or the matching distribution 224b. |A| is the absolute value of the difference between the pixel value Iv and the pixel value β.

In step S308, the absorbance ratio calculator 211 calculates the absorbance ratio 226 from the absorbance 225a at the target blood vessel cross-section position included in the image 221a and the absorbance 225b at the target blood vessel cross-section position included in the image 221b. Specifically, the absorbance ratio calculation unit 211 calculates the absorbance ratio 226 by Equation (2). OD _λ1 is the absorbance of the blood vessel at the cross-sectional position of the target blood vessel included in the image 221a captured by irradiating the eyeball 103 with light of the peak wavelength _λ1 . That is, OD _λ1 is the absorbance 225a. On the other hand, OD _λ2 is the absorbance of the blood vessel at the target blood vessel cross-sectional position included in the image 221b captured by irradiating the eyeball 103 with light of the peak wavelength _λ2 . That is, OD _λ2 is the absorbance 225b.

In step S309, the blood vessel contrast index calculation unit 212 calculates, as the blood vessel contrast index 227, a value having a correlation with the variance of the pixel values indicated by the matching distribution 224a and the peak height of the pixel values indicated by the matching distribution 224a. . Alternatively, in step S309, the blood vessel contrast index calculation unit 212 sets a value having a correlation between the variance of the pixel values indicated by the matching distribution 224b and the peak height of the pixel values indicated by the matching distribution 224b as the blood vessel contrast index 227. calculate. For example, the blood vessel contrast index calculation unit 212 calculates the blood vessel contrast index 227 using Equation (3). σ is a value indicating the variation from the average value of the pixel values indicated by the blood vessel cross-sectional profile 223a or the blood vessel cross-sectional profile 223b.

|A| represents the difference between the brightness of the inner center of the vessel wall and the brightness of the exterior of the vessel wall. Furthermore, 2σ indicates the blood vessel diameter at the target blood vessel cross-sectional position. Therefore, the blood vessel contrast index 227 according to the present embodiment is determined using factors such as the difference between the brightness at the center of the inside of the blood vessel wall and the brightness outside the blood vessel wall, and the blood vessel diameter at the cross-sectional position of the blood vessel.

In step S310, the distribution creation unit 213 determines whether or not the absorbance ratio 226 and the blood vessel contrast index 227 have been calculated for all blood vessel cross-sectional positions. If the absorbance ratio 226 and the blood vessel contrast index 227 have not been calculated for all blood vessel cross-sectional positions in step S310, the processing unit 204 returns the process to step S305. That is, the processing unit 204 repeats the processing from step S305 to step S310 until the absorbance ratio 226 and the blood vessel contrast index 227 are calculated for all blood vessel cross-sectional positions. On the other hand, if the absorbance ratio 226 and the blood vessel contrast index 227 have been calculated at all blood vessel cross-sectional positions in step S310, the processing unit 204 shifts the process to step S311.

In step S<b>311 , the distribution creation unit 213 creates a two-dimensional distribution 228 of the absorbance ratio 226 and the blood vessel contrast index 227 . Then, the processing unit 204 shifts the processing to step S801 illustrated in FIG.

Next, the processing performed by the measuring device 101 will be continued with reference to FIG.

In step S801, the blood vessel type estimation unit 214 clusters the two-dimensional distribution 228 to obtain clusters of a plurality of blood vessel types. For example, the blood vessel type estimation unit 214 clusters the two-dimensional distribution 228 using the k-means method, Gaussian mixture model, or the like.

In subsequent step S802, the blood vessel type estimation unit 214 estimates the blood vessel types at a plurality of blood vessel cross-sectional positions. Specifically, the blood vessel type estimation unit 214 identifies the cluster acquired in step S801 to which the set of the absorbance ratio 226 and the blood vessel contrast index 227 at the blood vessel cross-sectional position for which the blood vessel type is to be estimated belongs. , to estimate the type of blood vessel at the cross-sectional position of the blood vessel.

In step S803, the oxygen saturation calculator 215 calculates the oxygen saturation from the absorbance ratios 226 included in the groups belonging to a specific cluster.

Formula (4) is a calculation formula for calculating the oxygen saturation α from the absorbance ratio 226 . In the calculation formula shown in Equation (4), it is assumed that the light with peak wavelength λ ₁ and the light with peak wavelength λ ₂ have the same optical path length in the eyeball from when the light is emitted to when it is received by the imaging unit 201 . do. ε _Hb570 is the molar extinction coefficient of reduced hemoglobin for light with a peak wavelength λ ₂ =570 nm. _εHbλ1 is the molar extinction coefficient of reduced hemoglobin for light of peak wavelength _λ1 . _εHbO2λ1 is the molar extinction coefficient of oxygenated hemoglobin for light of peak wavelength _λ1 . Since ε _Hb570 , ε _Hbλ1 , and ε _HbO2λ1 are known constants, there is a first-order linear relationship between the oxygen saturation α and the absorbance ratio 226 as shown in Equation (4).

FIG. 9 is a diagram showing an example of the two-dimensional distribution 228. As shown in FIG. The two-dimensional distribution 228 shown in FIG. 9 has the absorbance ratio on the horizontal axis and the blood vessel contrast index on the vertical axis. Each circle shown in the two-dimensional distribution 228 illustrated in FIG. 9 represents a pair of absorbance ratio and blood vessel contrast index.

A frequency distribution 901 illustrated in FIG. 9 indicates the frequency of each set of absorbance ratios 226 included in the two-dimensional distribution 228 . In the frequency distribution 901, the horizontal axis indicates the absorbance ratio, and the vertical axis indicates the frequency. A frequency distribution 902 illustrated in FIG. 9 indicates the frequency of each set of blood vessel contrast index 227 included in the two-dimensional distribution 228 . In the frequency distribution 902, the horizontal axis is the frequency, and the vertical axis is the blood vessel contrast index.

The epiconjunctival artery and epiconjunctival vein exchange oxygen with oxygen in the air, resulting in a higher oxygen saturation than the episcleral artery and episcleral vein. The oxygen saturation of the episcleral artery is different from the oxygen saturation of the episcleral vein because arterial oxygen saturation is lower than venous oxygen saturation. As exemplified in equation (4), the absorbance ratio of the epiconjunctival artery and the epiconjunctival vein is the absorbance ratio of the episcleral artery, and the absorbance ratio of the episcleral artery, since there is a first-order linear relationship between the oxygen saturation and the absorbance ratio. It will be different from the absorbance ratio of the vein. Similarly, the absorbance ratio for the episcleral artery will be different than the absorbance ratio for the episcleral vein.

Therefore, a plurality of peak frequencies appear in the frequency distribution 901 of the absorbance ratio. For example, the absorbance ratio of one peak power out of the plurality of peak powers indicates the absorbance ratio of the epiconjunctival artery and the epiconjunctival vein. Also, for example, the absorbance ratio of other peak frequencies among the plurality of peak frequencies indicates the absorbance ratio of the episcleral vein.

Also, since the conjunctiva exists in a layer above the sclera, the supraconjunctival artery and supraconjunctival vein are imaged with a different clarity than the episcleral artery and episcleral vein in the images 221a and 221b. Therefore, the blood vessel contrast index of the blood vessel image of the epiconjunctival artery and epiconjunctival vein is calculated to have a different value from the blood vessel contrast index of the episcleral artery and episcleral vein. For example, the blood vessel contrast index of the blood vessel image of the epiconjunctival artery and epiconjunctival vein shows a higher value than the blood vessel contrast index of the episcleral artery and episcleral vein.

Therefore, a plurality of peak frequencies appear in the blood vessel contrast index frequency distribution 902 . For example, the vascular contrast index of one peak power out of the plurality of peak powers indicates the vascular contrast index of the episcleral artery and the vascular contrast index of the episcleral vein. Also, for example, other blood vessel contrast indicators of the plurality of peak frequencies indicate blood vessel contrast indicators of the epiconjunctival artery and epiconjunctival vein.

In this way, with respect to the epiconjunctival artery and epiconjunctival vein, the episcleral artery, and the episcleral vein, the absorbance ratio 226 and the blood vessel contrast index 227 are concentrated in different sections for each blood vessel type. Therefore, as illustrated in FIG. 9, the blood vessel type estimation unit 214 clusters the two-dimensional distribution 228 into a cluster of supraconjunctival arteries and supraconjunctival veins, a cluster of supraconjunctival arteries, and a cluster of supraconjunctival veins. . The blood vessel type estimation unit 214 identifies clusters representing the epiconjunctival artery and epiconjunctival vein, the episcleral artery, and the episcleral vein, and based on the absorbance ratio of the blood vessel cross-sectional position and the blood vessel contrast index, The type of blood vessel at the cross-sectional position of the blood vessel can be estimated.

Furthermore, the oxygen saturation calculator 215 can calculate the oxygen saturation of the episcleral artery from the absorbance ratios 226 included in the groups belonging to the cluster indicating the episcleral artery. Thereby, the measuring device 101 according to the present embodiment can estimate the oxygen saturation of the episcleral artery reflecting the health condition of the person without contact.

(Second embodiment)
A second embodiment will be described with reference to FIG. In the drawings, the same or similar elements are denoted by the same reference numerals, and overlapping descriptions are omitted.

Since the configuration of the measuring device 101 according to this embodiment is as shown in FIG. 2, detailed description thereof will be omitted.

The blood vessel contrast index calculation unit 212 according to the present embodiment calculates, as the blood vessel contrast index 227, a value having a correlation with the peak height of the pixel values indicated by the matching distribution 224a or the matching distribution 224b.

FIG. 10 is a flowchart showing an example of processing executed by the measuring device 101 according to this embodiment. The processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. 10 are the same as the processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. omitted.

In step S308, when the absorbance ratio calculation unit 211 calculates the absorbance ratio 226 from the absorbance 225a at the target blood vessel cross-section position included in the image 221a and the absorbance 225b at the target blood vessel cross-section position included in the image 221b, the blood vessel The contrast index calculator 212 calculates, as the blood vessel contrast index 227, a value having a correlation with the peak height of the pixel values indicated by the matching distribution 224a or the matching distribution 224b. For example, the blood vessel contrast index calculation unit 212 calculates the blood vessel contrast index 227 using Equation (5).

β is the pixel value on the baseline Io at the position of the minimum pixel value Iv in the matching distribution 224a or the matching distribution 224b. That is, β indicates the luminance outside the blood vessel wall. Also, A is a difference value between the pixel value Iv and the pixel value β. In other words, A indicates the brightness of the inner center of the blood vessel wall. |A| indicates the difference between the brightness of the inner center of the blood vessel wall and the brightness of the outside of the blood vessel wall. Therefore, the blood vessel contrast index 227 according to the present embodiment is determined using the brightness of the inner center of the blood vessel wall and the brightness of the outside of the blood vessel wall as factors, without depending on the blood vessel diameter. Therefore, the blood vessel contrast index calculation unit 212 according to the present embodiment can calculate the blood vessel contrast index 227 from the brightness of the inner center of the blood vessel wall and the brightness of the outside of the blood vessel wall for various blood vessels having different diameters.

Furthermore, the distribution generating unit 213 according to the present embodiment generates a blood vessel contrast index 227 calculated from the luminance of the center of the inside of the blood vessel at each blood vessel cross-section position and the luminance of the outside of the blood vessel wall, and the absorbance of the blood vessel at each blood vessel cross-section position. Create a two-dimensional distribution 228 with the ratio 226 . Then, the blood vessel type estimation unit 214 according to the present embodiment clusters the created two-dimensional distribution 228 to estimate the blood vessel types at a plurality of blood vessel cross-sectional positions.

As described above, the measurement apparatus 101 according to the present embodiment evaluates the clarity of the blood vessel image independently of the blood vessel diameter for various blood vessels having different blood vessel diameters, and then determines the type of blood vessel at a plurality of blood vessel cross-sectional positions. can be estimated.

(Third embodiment)
A third embodiment will be described with reference to FIG. In the drawings, the same or similar elements are denoted by the same reference numerals, and overlapping descriptions are omitted.

Since the configuration of the measuring device 101 according to this embodiment is as shown in FIG. 2, detailed description thereof will be omitted.

The blood vessel contrast index calculation unit 212 according to the present embodiment calculates a value having a correlation with the statistical value of the pixel values of a plurality of pixels at each blood vessel cross-sectional position indicated by the blood vessel cross-sectional profile 223a or the blood vessel cross-sectional profile 223b as a blood vessel contrast index. 227.

FIG. 11 is a flowchart showing an example of processing executed by the measuring device 101 according to this embodiment. The processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. 11 are the same as the processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. omitted.

In step S308, when the absorbance ratio calculation unit 211 calculates the absorbance ratio 226 from the absorbance 225a of the target blood vessel cross-section position included in the image 221a and the absorbance 225b of the target blood vessel cross-section position included in the image 221b, the blood vessel The contrast index calculator 212 calculates, as the blood vessel contrast index 227, a value having a correlation with the statistical values of the pixel values of the plurality of pixels at the target blood vessel cross-section position. For example, the blood vessel contrast index calculation unit 212 calculates the blood vessel contrast index 227 using Equation (6).

The blood vessel contrast index 227 calculated by Equation (6) indicates the average value of the amount of pixel value displacement between adjacent pixels at the blood vessel cross-sectional position to which the blood vessel cross-sectional profile 223a or the blood vessel cross-sectional profile 223b belongs. That is, the blood vessel contrast index 227 calculated by Equation (6) is a value that reflects changes in pixel values over one blood vessel cross-sectional position. Therefore, the blood vessel contrast index calculation unit 212 according to the present embodiment calculates the clarity around the blood vessel wall indicated by one blood vessel cross-sectional position and the clarity of the center of the blood vessel at one blood vessel cross-sectional position in the image 221a or image 221b. Equivalently, the blood vessel contrast index 227 can be calculated.

Therefore, the distribution generating unit 213 according to the present embodiment includes a blood vessel contrast index 227 that equally reflects the clarity around the blood vessel wall at each blood vessel cross-sectional position and the blood vessel center clarity at each blood vessel cross-sectional position, and A two-dimensional distribution 228 is created with the absorbance ratio 226 of the blood vessel at the cross-sectional position of the blood vessel. Then, the blood vessel type estimation unit 214 according to the present embodiment clusters the created two-dimensional distribution 228 to estimate the blood vessel types at a plurality of blood vessel cross-sectional positions.

As described above, the measurement apparatus 101 according to the present embodiment equally evaluates the clarity around the blood vessel wall indicated by one cross-sectional position of the blood vessel and the clarity of the center of the blood vessel at one cross-sectional position of the blood vessel, The types of blood vessels at multiple blood vessel cross-sectional positions can be estimated.

(Fourth embodiment)
A fourth embodiment will be described with reference to FIG. In the drawings, the same or similar elements are denoted by the same reference numerals, and overlapping descriptions are omitted.

Since the configuration of the measuring device 101 according to this embodiment is as shown in FIG. 2, detailed description thereof will be omitted.

The blood vessel contrast index calculation unit 212 according to the present embodiment calculates, as the blood vessel contrast index 227, a value having a correlation with the absorbance at each blood vessel cross-sectional position calculated for at least one of the images 221a and 221b. .

FIG. 12 is a flowchart showing an example of processing executed by the measuring device 101 according to this embodiment. The processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. 12 are the same as the processes of steps S301 to S308 and steps S310 to S311 illustrated in FIG. omitted.

In step S308, when the absorbance ratio calculation unit 211 calculates the absorbance ratio 226 from the absorbance 225a of the target blood vessel cross-section position included in the image 221a and the absorbance 225b of the target blood vessel cross-section position included in the image 221b, the blood vessel The contrast index calculation unit 212 calculates, as the blood vessel contrast index 227, a value having a correlation with the absorbance at the target blood vessel cross-sectional position calculated for at least one of the images 221a and 221b.

For example, the blood vessel contrast index calculator 212 calculates the sum of the absorbance 225a and the absorbance 225b as the blood vessel contrast index 227, as shown in Equation (7). OD _λ1 and OD _λ2 exemplified in Equation (7) are absorbance 225a and absorbance 225b, respectively.

Alternatively, the blood vessel contrast index calculator 212 may calculate the absorbance 225a or the absorbance 225b as the blood vessel contrast index 227 as shown in Equation (8).

Alternatively, the blood vessel contrast index calculator 212 may calculate the product of the absorbance 225a and the absorbance 225b as the blood vessel contrast index 227, as shown in Equation (9).

Therefore, the distribution generating unit 213 according to the present embodiment calculates the blood vessel contrast index 227 at each blood vessel cross-sectional position calculated from the absorbance of at least one of the images 221a and 221b, and the blood vessel contrast index 227 at each blood vessel cross-sectional position. A two-dimensional distribution 228 is created with the absorbance ratio 226 . Then, the blood vessel type estimation unit 214 according to the present embodiment clusters the created two-dimensional distribution 228 to estimate the blood vessel types at a plurality of blood vessel cross-sectional positions.

As described above, the measuring device 101 according to the present embodiment can estimate the blood vessel type from the blood vessel contrast index 227 and the absorbance ratio 226 calculated from the absorbance without contact.

The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments is also included in the technical scope of the present invention. Furthermore, new technical features can be formed by combining the technical means disclosed in each embodiment.

REFERENCE SIGNS LIST 100 measurement system 101 measurement device 102a light source 102b light source 103 eyeball 104a emitted light 104b emitted light 105a reflected light 105b reflected light 201 imaging unit 202 storage unit 203 control unit 204 processing unit 205 Image acquisition unit 206 Blood vessel region extraction unit 207 Blood vessel cross section determination unit 208 Blood vessel cross-section profile acquisition unit 209 Suitable distribution identification unit 210 Absorbance calculation unit 211 Absorption ratio calculation unit 212 Blood vessel contrast index calculation unit 213 Distribution creation 214 blood vessel type estimation unit 215 oxygen saturation calculation unit 221a image 221b image 222a blood vessel region 222b blood vessel region 223a blood vessel cross-sectional profile 223b blood vessel cross-sectional profile 224a matching distribution 224b matching distribution 225a absorbance 225b absorbance, 226 absorbance ratio, 227 blood vessel contrast index, 228 two-dimensional distribution, 601 blood vessel cross-sectional position, 901 frequency distribution, 902 frequency distribution

Claims (10)

an image acquisition unit that acquires at least two images including a blood vessel image of an eyeball captured by irradiation with at least two types of light having peak wavelengths different from each other;
a blood vessel cross-section profile acquisition unit configured to acquire a plurality of blood vessel cross-section profiles respectively indicating pixel value distributions of pixels at a plurality of blood vessel cross-section positions for each of the at least two images;
an absorbance calculation unit that calculates absorbance at each of the plurality of blood vessel cross-sectional positions from the plurality of blood vessel cross-sectional profiles;
creating a plurality of sets of two-dimensional distributions, each set including an absorbance ratio of each blood vessel cross-section position and a blood vessel contrast index indicating the clarity of each blood vessel cross-section position of the blood vessel image, wherein the absorbance ratio is at least a distribution generator calculated from the absorbance at each blood vessel cross-sectional position for the two images;
a blood vessel type estimation unit that clusters the two-dimensional distribution to obtain clusters of a plurality of types of blood vessels and estimates the types of blood vessels at the plurality of blood vessel cross-sectional positions;
A measuring device comprising: The measuring device according to claim 1, further comprising an oxygen saturation calculator that calculates oxygen saturation from absorbance ratios included in groups belonging to a specific cluster among the plurality of types of clusters. the plurality of types includes a type indicative of an episcleral artery;
3. The measuring device according to claim 2, wherein said specific cluster is a cluster of a type indicating said episcleral artery. Further comprising a blood vessel contrast index calculation unit for calculating the blood vessel contrast index from a blood vessel cross-section profile indicating a distribution of pixel values of pixels at each blood vessel cross-section position acquired for at least one of the at least two images. , the measuring device according to any one of claims 1 to 3. further comprising a matching distribution identification unit that identifies a matching distribution that matches the blood vessel cross-sectional profile;
5. The measuring apparatus according to claim 4, wherein said blood vessel contrast index calculation unit calculates, as said blood vessel contrast index, a value having a correlation with a peak of pixel values indicated by said suitable distribution. the fitted distribution is a normal distribution;
6. The measuring apparatus according to claim 5, wherein said blood vessel contrast index calculation unit calculates a value having a correlation with variance and peak height of said adaptive distribution as said blood vessel contrast index. The blood vessel contrast index calculation unit calculates each blood vessel cross section represented by a blood vessel cross section profile showing a distribution of pixel values of a plurality of pixels at each blood vessel cross section position acquired for at least one of the at least two images. 5. The measuring device according to claim 4, wherein a value having a correlation with statistical values of pixel values of a plurality of pixels at a position is calculated as said blood vessel contrast index. 5. The blood vessel contrast index calculation unit calculates, as the blood vessel contrast index, a value having a correlation with the absorbance of each of the blood vessel cross-sectional positions calculated for at least one of the at least two images. The measuring device according to . acquiring at least two images including a blood vessel image of an eyeball captured by irradiating at least two types of light having peak wavelengths different from each other;
acquiring a plurality of blood vessel cross-section profiles each showing a distribution of pixel values of pixels at a plurality of blood vessel cross-section positions for each of the at least two images;
calculating the absorbance at each of the plurality of blood vessel cross-sectional positions from the plurality of blood vessel cross-sectional profiles;
creating a plurality of sets of two-dimensional distributions, each set including an absorbance ratio of each blood vessel cross-section position and a blood vessel contrast index indicating the clarity of each blood vessel cross-section position of the blood vessel image, wherein the absorbance ratio is at least calculating from the absorbance at each blood vessel cross-sectional position for the two images;
a step of clustering the two-dimensional distribution to obtain clusters of a plurality of blood vessel types, and estimating the blood vessel types at the plurality of blood vessel cross-sectional positions;
including measurement method. to the computer,
a function of acquiring at least two images including a blood vessel image of an eyeball captured by irradiating at least two types of light having different peak wavelengths;
a function of acquiring a plurality of blood vessel cross-section profiles respectively showing pixel value distributions of pixels at a plurality of blood vessel cross-section positions for each of the at least two images;
a function of calculating the absorbance at each of the plurality of blood vessel cross-sectional positions from the plurality of blood vessel cross-sectional profiles;
creating a plurality of sets of two-dimensional distributions, each set including an absorbance ratio of each blood vessel cross-section position and a blood vessel contrast index indicating the clarity of each blood vessel cross-section position of the blood vessel image, wherein the absorbance ratio is at least a function calculated from the absorbance at each blood vessel cross-sectional position for the two images;
a function of clustering the two-dimensional distribution to obtain clusters of a plurality of blood vessel types, and estimating the blood vessel types at the plurality of blood vessel cross-sectional positions;
program to run.

Images