Analyte measuring devices and related methods

Abstract

A technology is described for detecting a target analyte in skin of a subject. The method can include selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to a wavelength absorbed by the target analyte. Light may be emitted from the one or more LEDs toward a subject's skin. A light intensity value of the light emitted by the one or more LEDs may be measured with at least one photodiode before the light's interaction with the subject's skin. A light intensity value of the light emitted by the one or more LEDs may be measured measuring with at least one photodiode after the light's interaction with the subject's skin. Another operation is providing a correlation output for measurements of the light using a machine learning model.

Description

Thorpe North & Western, LLP Docket No.: 01336-36037.PCTANALYTE MEASURING DEVICES AND RELATED METHODSPRIORITY DATA

[0001] This application claims the benefit of United States Provisional Patent Application Serial No. 63 / 748,911, filed on January 23, 2025, which is incorporated herein by reference.BACKGROUND

[0002] Reactive oxygen species may be a part of a living organisms’ biological defense mechanisms. However, excessive reactive oxygen species in a body may lead to various diseases or disorders. Humans have a series of antioxidant defense systems to protect against reactive oxygen species. For antioxidant systems to operate normally, it is useful to have adequate amounts of both endogenously produced and exogenously administered antioxidants such as vitamin C, vitamin D, carotenoids, flavonoids, or tocopherol.

[0003] Raman spectroscopy has been used for the measurement of carotenoids in living tissues. Raman spectroscopy uses a focused monochromatic light source (e.g., blue light) to excite carotenoids, allowing these tissue antioxidants to be characterized by their specific vibrational energy levels. This method has been validated through many scientific studies. Biophotonic scanners have used Resonant Raman Spectroscopy to quantify carotenoids in intact human skin as an indicator of nutritional intake and in vivo antioxidant status.BRIEF DESCRIPTION OF THE DRAWINGS

[0004] FIG. 1 is a perspective view of an example reflection spectroscopy (RS) device (analyte measuring device) that may be used to measure target analytes in the skin.

[0005] FIG. 2 is a block diagram illustrating a cross-sectional view of an example reflection spectroscopy device (analyte measuring device) capable of detecting a target analyte in skin of a subject.

[0006] FIGs. 3A-3K illustrate an example assembly process for an analyte measuring device.

[0007] FIG. 3L illustrates an example analyte measuring system.

[0008] FIG. 4 is a chart illustrating an example of a single scan cycle for the scanning device.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT

[0009] FIG. 5 is a chart illustrating an example definition of part of a total scan that includes multiple scan cycles.

[0010] FIG. 6 is a chart illustrating another example definition of part of a total scan that includes multiple scan cycles.

[0011] FIG. 7 is a chart illustrating another example definition of a scan that includes multiple scan cycles.

[0012] FIG. 8 a flowchart illustrating a method of detecting a target analyte in skin of a subject.

[0013] FIG. 9 is flowchart illustrating an example of a method for providing a decision output from machine learning related to tissue of an individual using a spectral light sensor.

[0014] FIG. 10 is flowchart illustrating an example of a method for scanning tissue of a subject and providing correlation output from a machine learning model.

[0015] FIG. 11 is a flow chart illustrating an example of a method of training a machine learning model to provide a prediction for a target analyte in an area of a subject’s skin.

[0016] FIGs. 12A and 12B are graphs of S3 scanner score vs. Prysm scanner score.

[0017] FIGs. 13A-13D are graphs of Pr sm scanner scores and S3 scanner scores of several groups in a study over time.

[0018] FIGs. 14A and 14B graphs of serum concentration of carotenoids vs. Prysm scanner score.

[0019] FIG. 14C is a graph of Prysm scanner score vs. S3 scanner score.DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Although Raman spectroscopy -based scanners can be used to detect carotenoids in human skin, these devices present a number of limitations. The technology described herein provides improved devices that can assess the compositional state of tissue, including the content of desirable and / or undesirable analytes and that can be used to identify and model related health conditions. Compared to Raman spectroscopy-based devices, the analyte measuring devices described herein can be easier to use, more cost-effective, and more energy efficient.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT Definitions

[0021] In describing and claiming the present invention, the following terminology will be used.

[0022] The singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a photodiode” includes reference to one or more of such features and reference to “activating” refers to one or more of such steps.

[0023] As used herein with respect to an identified property or circumstance, “substantially” refers to a degree of deviation that is sufficiently small so as to not measurably detract from the identified property or circumstance. The exact degree of deviation allowable may in some cases depend on the specific context.

[0024] As used herein, the term “about” is used to provide flexibility and imprecision associated with a given term, metric or value. The degree of flexibility for a particular variable can be readily determined by one skilled in the art. However, unless otherwise enunciated, the term “about” generally connotes flexibility of less than 2%. and most often less than 1%, and in some cases less than 0.01%.

[0025] As used herein, a plurality of items, structural elements, compositional elements, and / or materials may be presented in a common list for convenience. However, these lists should be construed as though each member of the list is individually identified as a separate and unique member. Thus, no individual member of such list should be construed as a de facto equivalent of any other member of the same list solely based on their presentation in a common group without indications to the contrary'.

[0026] As used herein, the term “at least one of’ is intended to be synonymous with “one or more of.” For example, “at least one of A. B and C” explicitly includes only A, only B, only C, or combinations of each.

[0027] Numerical data may be presented herein in a range format. It is to be understood that such range format is used merely for convenience and brevity and should be interpreted flexibly to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. For example, a numerical range of about 1 to about 4.5 should be interpreted to include not only the explicitly recited limits of 1 to about 4.5, but also to include individual numerals such as 2, 3, 4, and sub-ranges such as I to 3, 2 to 4, etc. The same principle applies to ranges reciting only one numerical value, such as “less than about 4.5,” which should be interpreted toThorpe North & Western, LLP Docket No.: 01336-36037.PCT include all of the above-recited values and ranges. Further, such an interpretation should apply regardless of the breadth of the range or the characteristic being described.

[0028] Any steps recited in any method or process claims may be executed in any order and are not limited to the order presented in the claims. Means -plus -function or stepplus-function limitations will only be employed where for a specific claim limitation all of the following conditions are present in that limitation: a) "‘means for” or “step for” is expressly recited; and b) a corresponding function is expressly recited. The structure, material or acts that support the means-plus function are expressly recited in the description herein. Accordingly, the scope of the invention should be determined solely by the appended claims and their legal equivalents, rather than by the descriptions and examples given herein.Example Embodiments

[0029] Reference will now be made to the examples illustrated in the drawings, and specific language will be used herein to describe the same. It will nevertheless be understood that no limitation of the scope of the technology is thereby intended. Alterations and further modifications of the features illustrated herein, and additional applications of the examples as illustrated herein, which would occur to one skilled in the relevant art and having possession of this disclosure, are to be considered within the scope of the description.

[0030] Consumers are looking for ways to assess wellness and reinforce positive lifestyle choices in their everyday lives. Devices and systems can be used that provide non-invasive methods of detecting target analytes (e.g. antioxidants) in biological tissue, such as skin and provide data that allows assessment (e.g. quantification and / or measurement) of the impact of diet, supplements or other health factors on physiologic status and uality of life (QOL). Such status can then be used in forming or developing dietary, supplement, and / or activity regimens designed to correct noted deficiencies or otherwise optimize various chemical and / or biochemical species within a subject and thereby maximize health and wellbeing.

[0031] FIG. 1 illustrates a perspective view of a reflection or reflectance spectroscopy (RS) device 110 that may be used to measure analytes, such as antioxidants, in the skin or tissue of a subject. Such devices may be used for providing personalized health guidelines indicating or otherwise indicative of or correlating to, an individual's characteristics, including antioxidants, for health management of the individual. For example, a subject may put an area of skin to be tested (e.g. a digit such as a finger) on an aperture 112 of the device and light may interact with the user’s skin and this interaction (e.g., reflected light from the skin or light entering into or otherwise encountering or engaging the skin and returning fromThorpe North & Western, LLP Docket No.: 01336-36037.PCT therefrom) may be used for or otherwise result in measurements representing one or more target analytes, such as antioxidants in the skin of the user. This example device also includes a power button 114 and a charging port 116.

[0032] In some embodiments, the device may use hyperspectral imaging or absorption. For example, one or more LEDs (e.g. an array) can be mounted inside the device, for example, on a printed circuit board (PCB) inside the device. Light can be emitted from the one or more LEDs and pass through the aperture 112 (e.g. a window) in the housing of the device, against which an area of skin, tissue, or another sample (such as a calibration standard sample) to be analyzed is applied. The individual LEDs can emit a single wavelength or color of light, or in some cases the individual LEDs can emit more than one wavelength or color of light. Additionally, different LEDs can emit different wavelengths or colors compared to the other LEDs in some examples, while in some examples multiple LEDs may emit the same wavelengths or colors. After passing through the window, the light can interact with the area of skin, tissue, or other sample. Analytes present in the skin can absorb a portion of the light energy while the remaining light energy (e.g. unabsorbed light energy’) can reflect off the skin back through the window toward one or more light sensors. For example, a plurality of light sensors, such as photodiodes, can be included within the device (e.g. mounted on a PCB. One or more of the photodiodes can measure light characteristics, such as brightness or intensity7of outgoing light emitted from the one or more LEDs (e.g., before or immediately before passing through the window or after passing through the window and before interacting with the skin) and another photodiode can measure light characteristics (e.g. brightness, wavelength, color, or a combination of these) of the light after interacting with the subject's tissue or skin. These measurements allow the device to measure how much light is reflected and / or absorbed by the tissue and can be quantified in several ways, including as a ratio value that is subsequently used by one or more machine learning systems. In alternative embodiments, the amount or other characteristics of light emitted from the one or more LEDs can be inferred or otherw ise derived from the characteristics and properties of the at least one, or combination of, LEDs instead of, or in combination with measuring its output via one or more sensors. In one embodiment, the device may include an array of five LEDs which activate and emit light in sequence, then have one ALL OFF period in a cycle. The device may cycle for prescribed or predetermined period of time, such as about 15 seconds, and for a prescribed or predetermine number of cycles, such as at least 50 cycles, or at least 100 cycles, or at least 300 cycles, which may generate a large number (e.g. 190,000) of measurements.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT Based on the measurements made, the device derives a final output or measurement value for the user.

[0033] In some embodiments, the RS device 110 aperture may be of sufficient size that it is substantially or entirely occluded by an appendage of a subject or user (e.g. a finger / fingertip) when placed thereover. Such occlusion can prevent entrance of ambient light into a measurement area of the RS device and improve or maximize accuracy of measurements taken. In some other embodiments, the aperture can be less than substantially or entirely occluded when used by a subject and the RS device can be calibrated or otherwise designed or programmed to compensate for entry of ambient light into the measurement area within the device.

[0034] FIG. 2 is a block diagram illustrating example components of a RS device 200 capable of detecting a target analyte in skin of a subject. One or more light emitting diodes (LEDs) 206a-e can be disposed within a housing 202 of the RS device and positioned to direct light toward an aperture 208. The one or more light emitting diodes (LEDs) may be selected to emit light, and the LEDs may have one or more wavelengths corresponding to a wavelength absorbed by the target analyte. As shown in FIG. 2, the one or more LEDs are positioned on a printed circuit board (PCB) 204 housed within the housing 202. However, in other examples the device is not limited to this specific arrangement and the LEDs can be placed at any location within the RS device that is suitable for operation or that provides a particular characteristic or advantage. In some embodiments, the LEDs may be arranged away or apart from the PCB.

[0035] In use, light may be emitted from the one or more (e.g. plurality' of) light emitting diodes (LEDs) 206a-e toward a subject’s skin or tissue. The light may pass through the aperture 208 (and / or window 220) in the reflection spectroscopy device 200 or housing and interact with a subject’s tissue. For example, the light may interact with an analyte in the subject’s tissue and then be reflected back. In some embodiments, the target analyte may be at least one of: a vitamin, a mineral, a pigment (e.g., melanin), a lipid, water, a protein (e.g. collagen or elastin), oxygen, hemoglobin, etc. Further, the target analyte may more specifically be an antioxidant that is at least one of: carotenoids, vitamin A, vitamin C, vitamin D, vitamin E, polyphenols, zinc, or selenium. In addition, the tissue may be a person’s skin, an animal's skin or the tissue of a plant.

[0036] In some embodiments, the one or more LEDs may have a wavelength of between 450 nm and 690 nm. In one example, the one or more LEDs includes, or is exactly, 5 LEDs 206a-c, as illustrated in FIG. 2. Different target analytes may absorb light at differentThorpe North & Western, LLP Docket No.: 01336-36037.PCT wavelengths. The wavelengths that are selected may be known to be absorbed by a given analyte, such as a carotenoid. Other wavelengths may be absorbed by other analytes such as pigment (e.g., melanin) and hemoglobin, which are desired to be include or excluded from the final measurement. For instance, a final calculation may be adjusted to account for levels of pigment (e.g., melanin) in the subject’s skin. If light hits or otherwise encounters an analyte that absorbs the light, then the light bounces back or reflects at a weaker intensity. Generally speaking, the stronger and the more absorbent the analyte (e.g. the type or concentration of the analyte), the greater the reduction in the light that bounces off of, or otherwise returns from the skin (e.g. reflects off the skin). The number of LEDs, as w ell as the properties and characteristic of each, such as wavelengths or intensity of light emitted from each may be varied depending on the number and types of analytes that are desired to be scanned and measured.

[0037] A first light intensity value of the light emitted by the one or more LEDs may be measured with at least one first photodiode 210 before the light’s interaction with the subject's skin. Alternatively, these values can be derived or inferred from the specifications and operational parameters of the one or more LEDs either alone or in combination with the photodiode measurement. This first light intensity value may be used as a comparison value for a second light measurement that will be taken.

[0038] A second light intensity’ value of the light emitted by the one or more LEDs may be measured with at least one second photodiode 212 after the light’s interaction with the subject’s skin. Optionally, a number of specific structures can be included in the RS device 200 to separate the LEDs from the photodiode(s), thereby preventing direct exposure of a given photodiode to light emitted by an LED and improving accuracy of measurements. One example is a separation panel 222 which may be situated between the one or more LEDs 206a-e and the photodiode second 212 for reading light that is returning from the skin so that light leaving the LEDs is not directed to the second photodiode 212 measuring the light that is reflected off the skin. In other words, the first photodiode 210 measures how bright the light is on the way up to the skin while the second photodiode 212 measures the light that is reflected off the skin. In some examples, an additional separation panel or similar structure can be between the aperture and the first photodiode 210 to prevent the first photodiode from receiving light reflected back from the skin.

[0039] In a further example, the first photodiode 210 is positioned to sense light from the five LED lights when the lights are on the way up, (e.g. emitted toward the aperture 208) while the second photodiode 212 is positioned to sense light from the five LED lights as theThorpe North & Western, LLP Docket No.: 01336-36037.PCT light returns from the skin. It is also possible to have one photodiode for each LED for the light going up and / or coming back (e.g.. 10 photodiodes or 5 photodiodes), but this may depend on the sensing application or the size of the target tissue being sensed (e g., more photodiodes could be used for a larger tissue sensing area).

[0040] A correlation output may be provided using an output screen 214 for measurements of the light using a machine learning model. The correlation output may be a numeric value, an alpha-numeric value, or a classification tag provided as output from a machine learning model. This example device also includes a processor 216 electronically connected to the photodiodes through the PCB 204. The processor can be programmed to provide the correlation output based on the measurements provided by the photodiodes. The processors can be programmed to use the machine learning model to calculate the correlation output and then to display the correlation output on the output screen.

[0041] In one configuration, the light from the one or more LEDs may be passed through a first polarizing filter. The polarizing filter can be a layer formed or placed directly on the LEDs themselves, or a separate component that is placed between the LEDs and the aperture. Then the light may interact with the tissue or skin of the person. A portion of the light can be reflected off the surface the skin, while another portion of the light may penetrate into the skin. The portion of light reflected off the surface of the skin can retain the same polarization provided by the polarization filter. However, the light that penetrates the skin can be scattered by tissue under the surface of the skin. The scattering causes the light to become depolarized. The depolarized light can then be redirected back through the aperture. When the light has returned from the skin, the light that has interacted with the skin may be detected using the second photodiode 212. The light entering the second photodiode may pass through a second polarizing filter with a different orientation than the first polarizing filter. This second polarization filter can block the light that reflected off the surface of the skin and retained the original polarization, while allowing the depolarized light that penetrated the skin to pass through. Thus, the second photodiode can preferentially measure light that has penetrated into the skin. This can be useful because the light that penetrated the skin has had more opportunity to be absorbed by target analytes, and therefore can provide more information about the target analytes present in the skin. In one example, measuring light may occur both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the skin using an emission measuring photodiode, and 2) as the light returns from the skin using an interaction measuring photodiode.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT

[0042] Notably, the one or more polarizing filters can be disposed at any suitable location within the RS device 200. In some embodiments, such polarizing filters can be located between the one or more LEDs 206a-e and the aperture 208. In one embodiment, at least one polarizing filter can be immediately proximate one or more of the LEDs. For example, a polarizing film can be applied directly over the LEDs. In another embodiment, at least one polarizing filter can be located proximate one or more photodiodes. For example, a polarizing film can be applied directly over the photodiode. In some embodiments, the polarizing filters can be substantially the same and in other embodiments, they can be different.

[0043] An exploded view of another example analyte measuring device (also referred to as an RS device as explained above) is shown in FIG. 3A. This device 300 includes a bottom housing 302 with a power button 304 shown as a separate part that fits into the bottom housing. A PCB 306 is set into the bottom housing, with a battery’ 308 connecting to the PCB by a cable 310. The PCB includes an array of five LEDs 312 and photodiode 314. An optical chassis 320 is placed over the PCB. The optical chassis can direct light from the LEDs upward toward a subject’s skin. The optical chassis can also block out extraneous light so that unwanted ambient light is not received by the photodiode. The optical chassis can also include a separation panel to block the interaction measuring photodiode from receiving light directly from the LEDs. A cover 330 is placed over the optical chassis. The cover includes an aperture opening 332 to allow light from the LEDs to pass through the aperture of the device to the subject’s skin. The cover also includes a photodiode opening 334 that can allow light from the LEDs to pass directly to an emission measuring photodiode. A second PCB 340 is placed over the cover. The second PCB includes an array of indicator LEDs 342. An emission measuring photodiode can be located on the bottom of the second PCB, which is not visible in this view. The emission measuring photodiode can be aligned with the photodiode opening in the cover, so that light from the LEDs is received directly by the emission measuring photodiode. In this example, the various parts of the device are assembled using screws 350.

[0044] More detailed illustrations of the parts of this example analyte measuring device 300 are shown in FIGs. 3B-3L. 3B shows the bottom housing 302. Four magnets 352 are inserted into the bottom housing. The PCB 306 is connected to the battery’ 308 using the cable 310. A communication cable 354 is also connected to the PCB. This assembly is then set in the bottom housing. FIG. 3C shows a perspective view of these parts after assembly, and FIG. 3D shows a top plan view of these assembled parts. This view shows the array ofThorpe North & Western, LLP Docket No.: 01336-36037.PCT LEDs 312 and the photodiode 314 on the PCB. This photodiode operates as the interaction measuring photodiode, because this photodiode receives light that has interacted with the subject’s skin. The PCB can also include a processor 356 and other electronic components. In FIG. 3E, the power button 304 and optical chassis 320 are added using screws 350. FIG. 3F shows a perspective view of the assembled parts after this step. FIG. 3G shows a top plan view of these parts. The optical chassis includes a light chamber 322 that directs light from the LEDs upward to the aperture and allows light to be reflected from the subject’s skin back to the interaction measuring photodiode. A separation panel 324 is positioned between the array of LEDs and the interaction measuring photodiode. This separation panel only extends partially up from the bottom optical chassis. The height of the separation panel is sufficient to block light from travelling directly from the LEDs to the interaction measunng photodiode, while also allowing reflected light from the subject’s skin to travel back through the aperture and to the interaction measuring photodiode. FIG. 3H shows the cover 330, which is placed over the light chamber. The cover includes an aperture opening 332 that allows the light from the LEDs to travel up through the aperture to the subject’s skin. The aperture opening has a slanted wall one side to allow light to travel on a slanted angle back from the aperture toward the interaction measuring photodiode. The opposite wall of the aperture opening is vertical, which can help block reflected light from reaching the emission measuring photodiode. The emission measuring photodiode can be on the bottom surface of the second PCB 340 located to align with the photodiode opening 334 in the cover. The second PCB also includes an array of indicator LEDs 342 on the upper surface of the second PCB. FIG. 31 shows a perspective view of the assembled parts after this step. FIG. 3J shows a top plan view of these assembled parts. This view shows that the array of LEDs is located directly under the aperture. A top housing can be attached to cover the top parts of the device. FIG. 3K shows the device with the top housing 360 attached. The top housing includes a transparent window 362 over the aperture. A subject can place a finger or other skin over this window to scan the skin.

[0045] The aperture in the center of the top housing can be an opening or a window. A window can be transparent, semi-transparent, translucent, etc., to allow light from the LEDs to pass through the window to a subject’s finger, and to allow reflected light to pass back through the window to the interaction measuring photodiode.

[0046] The indicator LEDs can be used for convenience of the user, and are not used to provide the light for measuring the target analyte. Instead, the indicator LEDs can be used to indicate the progress of a scan in some examples. For example, if a scan runs for 15Thorpe North & Western, LLP Docket No.: 01336-36037.PCT seconds, then the indicator LEDs can be lit one by one throughout the 15 seconds until all indicator LEDs are lit when the scan is complete. The indicator LEDs can also be used to display the correlation output. For example, the indicator LEDs can operate as a display dial, in which a number of the LEDs can be lit to indicate a correlation output score. In further examples, the device can include an electronic display that can display the correlation output instead of, or in addition to, using the indicator LEDs.

[0047] In still further examples, the analyte measuring device can connect wirelessly to an external device, such as a mobile device, smartphone, tablet, personal computer, server, cloud network, and so on. In a particular example, the external device can display the correlation output. In certain examples, the correlation output can be a score that is based on the amount of the target analyte present in the subjects skin. The score can be a numeric score, alphanumeric score, or classification (such as classifying the subject into one of several different classes based on the amount of the target analyte detected). Any of these can be displayed using a display of the external device. The analyte measuring device can include a wireless communication module, such as a Bluetooth module, to allow the analyte measuring device to communicate wirelessly with the external device. In further examples, a wired connection can also be used to connect to an external device.

[0048] An analyte measuring device can also be a part of a sy stem. FIG. 3L shows one example system that includes an analyte measuring device 300, or skin scanning device, as described in the previous figures. The system also includes a correlation module 370. The correlation module can be implemented as software, hardware, or any combination thereof. The correlation module is in electronic communication with the skin scanning device such that the correlation module can receive photodiode measurements from the photodiodes of the skin scanning device. The correlation module can be programmed to provide the correlation output based on the measurements using a machine learning model. In some examples, the correlation module can include programming in the skin scanning device itself. For example, the skin scanning device can include a processor that is programmed to perform the calculations to provide the correlation output without requiring any external devices. In other examples, the correlation module can include software or hardware of one or more external devices. The one or more external devices can be programmed to calculate the correlation output. In the example shown in FIG. 3L, the correlation module includes a personal computer 372, a mobile device 374, a cloud network 376, and a server 378. In various examples, any of these external devices can be used individually or in any combination to calculate the correlation output. Any of these devices can be in electronicThorpe North & Western, LLP Docket No.: 01336-36037.PCT communication with each other. In some examples, the correlation module can allow the machine learning model to be updated. For example, an updated machine learning model can increase the accuracy of the correlation output. Updated machine learning models can be prepared by using additional training data and / or different machine learning programming. The correlation module can also utilize the photodiode measurements provided by the skin scanning device as training data for developing updated machine learning models in some cases.

[0049] The example analyte measuring devices illustrated above can have an aperture sized and shaped to be covered by a finger or thumb of the subject. In some examples, the aperture can include a solid transparent window and the subject can place a finger or thumb against the solid transparent window. In other examples, the aperture can include an opening and the subject can place the thumb or finger over the opening. The aperture can have an area smaller than the area of a finger or thumb. In certain examples, the aperture can have an area from about 25 mm2to about 200 mm2or from about 25 mm2to about 100 mm2. In alternative examples, the aperture can have a larger area. The aperture may also be designed to scan another skin area of the subject other than a finger or thumb. For example, the aperture can be sized and shaped to scan a palm, an arm, leg, trunk, forehead, face, neck, or other skin area of the subject. In some examples, the analyte measuring device can be in direct contact with the skin of the subject during the scan. In other examples, the device can be placed at a distance away from the skin, where the distance can be from about 1 mm to about 1 m, or from about 1 mm to about 10 cm, or from about 1 mm to about 2 cm. In further examples, the analyte measuring device can be configured to scan the skin while the skin is stationary with respect to the device. In alternative examples, the analyte measuring device can be configured to scan the skin while the device or the skin is moving with respect to the other. For example, the device can be configured to slide or roll across an area of skin while scanning.

[0050] In more detail regarding the LEDs in the scanning device, the use of the LEDs and detection of an analyte may occur in cycles. Each of the one or more LEDs may be activated to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle. Further, the cycle may be repeated at least 50 times, or at least 100 times, or at least 150 times, or at least 180 times, or at least 300 times, or another set number of times.

[0051] FIGs. 4 and 5 illustrate further details about the one or more LEDs and photodiode scanning and timing. An example of the light emitting diode (LED) and photodiode (PD) timing during a measurement using the reflection spectroscopy device willThorpe North & Western, LLP Docket No.: 01336-36037.PCT now be discussed. In this example, a single skin measurement may take 14.8 seconds. A full scan may be composed of 300 cycles. Of course, the scans may be longer or shorter or use more or fewer cycles.

[0052] This example in FIG. 4 shows the LED timing for each of five LEDs, and when each of the two PD measurements occur. The example device may contain five LEDs (470nm, 495nm, 520nm. 590nm and 630nm). The device illuminates only one LED at a given time. As a result, the example device may utilize six LED states to perform the measurement. These states may include an ALL-OFF state 410. So, the states may be ALL OFF 410, 470nm only, 495nm only, 520nm only, 590nm only and 630nm only 412. In an alternative embodiment, the LEDs may emit UV (ultraviolet) or IR (infrared) frequencies.

[0053] In one example of the technology, the brighter the LED lights, the more reliable results may be obtained. However, dimmer LEDs can provide acceptable results. The LEDs may be strong enough to provide ~5mW on the way up, and ~5mW when the light returns from the subject’s skin (e.g., on the way down). The LEDs light intensity may be varied, if desired, for some analyte detections.

[0054] In the example device, at least two photodiodes may be used. A photodiode is a semiconductor device that converts photons (or light) into electrical current. There may be two photodiodes in the optics block. The first photodiode (PD1) 416 may be positioned to measure the brightness of the LEDs. The second photodiode (PD0) 414 may be positioned to measure the amount of light reflected from the skin and back to the optics block. In some examples, the photodiodes can have an area from about 2 mm2to about 25 mm2, or from about 5 mm2to about 25 mm2.

[0055] An analog-to-digital converter can be used to obtain digital readings from the photodiodes. For example, the device may utilize a Nordic 14-bit analog-to-digital converter. The A / D converter may read the voltage from each photodiode for a time period, such as from about 0.01 ms to about 0.05 ms, or about 0.01 ms, or about 0.02 ms, or about 0.03 ms, or about 0.04 ms (milliseconds). Due to the short duration for the readout from the A / D chip, the voltage from the photodiode may be measured multiple times. The multiple readings may¬ be processed by a function to arrive at the final result. For example, an average of the multiple readings may be computed. Alternatively, some other function may be used for combining the readings such as using the median, mode, geometric average, harmonic average or another function. The photodiode may be read a certain number of times to obtain a final result, such as from about 50 times to about 300 times, or 65 times in some examples, or 256 times in some examples. The photodiode may be read any number of times to get theThorpe North & Western, LLP Docket No.: 01336-36037.PCT result. Therefore, in this example, each photodiode reading process may take approximately 2.6 ms (65 x 0.04 ms), or about 2.9 ms (256 x 0.01 ms), or other time depending on the time of each read and the number of reads. Alternatively, it is possible that the device may use just one photodiode for measuring light returning from the skin and use reference measurements in a data store of measurements for the light from the one or more LEDs.

[0056] A single scan cycle will now be discussed for this example device, as illustrated in the chart of FIG. 4. During a single scan cycle, the six LED States (including the ALL OFF state) are active for 8.2 ms each. This results in a total cycle time of 49.2 ms (6 x 8.2 ms) as depicted in FIG. 4. Each LED state may constitute the following three actions (in order):1. 3 ms pre-measurement delay.2. 65 measurements of the second photodiode (PD0), each lasting 0.04 ms. This results in 2.6 ms of measurement time for the second photodiode (PD0) (i.e. 65 x 0.04 ms = 2.6 ms).3. 65 measurements of the first photodiode (PD1), each lasting 0.04 ms. This results in 2.6 ms of measurement time for the first photodiode (PD1) (i.e. 65 x 0.04 ms = 2.6 ms).FIG. 4 illustrates this single cycle timing.

[0057] FIG. 5 is an example chart illustrating the definition of part of a total scan that includes three single cycles 502a-c. At the completion of a cycle, the next cycle will begin without an additional delay, as illustrated in FIG. 5. There may be 300 cycles in a scan, resulting in atotal scan time of 14.76 seconds (300 x 49.2 ms = 14.76 s). FIG. 5 illustrates three cycles performed without a cycle-to-cycle delay.

[0058] Using an approximately 15 second scan appears to provide accurate results. A time period longer than 15 seconds may be used but does not appear to be more accurate. A time period shorter than 15 seconds may be used but may be less accurate unless a significant number of samples is obtained. For example, the measurement(s) may be done in 1 to 3 seconds but this may affect how accurate the measurement is. The use of 300 cycles uses a selected processor at a high processing capacity in the 15 second scan period. However, other more powerful or less powerful processors may be used to scan faster or slower in the same window of time.

[0059] In some embodiments, the order in which the LEDs fire or emit light does not matter as long as the LED that is firing does so after the prior LED and after the prior LED has completely turned off so there is no interference or residual light remaining during eachThorpe North & Western, LLP Docket No.: 01336-36037.PCT cycle. In other embodiments, the order of LED illumination can be deliberately selected (or deliberately made random) for a number of reasons, such as achieving a desired result, doing quality control or calibration evaluation, in view of a specific target analyte known to be more accurately measured in such a way, etc. In some aspects, the timing set in the cycles can be used to increase the confidence that the previous LED light is completely gone or absent before the next LED turns on and is measured. In some embodiments, the photodiodes do not have knowledge about which LED is firing because the photodiodes just measure a signal inside the device. The software knows which LED is firing but the photodiodes do not utilize that information. In other embodiments, the photodiodes are aware of which LED is being illuminated.

[0060] Another example analyte measuring device can be configured to use four different colored LEDs. The colored LEDs can emit cyan, red, yellow, and green light, respectively. Each LED can be activated for a period of 13.49 ms. The LEDs can be activated individually such that two different LEDs are never activated at the same time. Thus, the device can cycle through four color states in which the LED of each color is turned on. The device can also have an all-off state in which none of the LEDs are activated. During each state, the device can be programmed to measure light intensity with two photodiodes. The first photodiode (the emission measuring photodiode) is labelled PD1. The second photodiode (the interaction measuring photodiode) is labelled PDO. During each color state and in the all-off state, the device is programmed to wait for a delay time, then measure the intensity of PD1, then measure the intensity of PDO, then re-measure the intensity of PDO and PD1. FIG. 6 shows a chart illustrating this cycle. The cycle is repeated 156 times in this example to complete a scan of the subject’s skin.

[0061] FIG. 7 shows a chart of another example schedule for activating LEDs and measuring the light intensity using the photodiodes. This example uses a different period of 9.7 ms for activating each LED and a 9.7 ms all-off state. This schedule includes performing 180 cycles in which the device performs a delay followed by measuring with PD1 and then PDO. After this, 180 more cycles are performed utilizing a delay followed by measuring with PDO and then PD1.

[0062] FIG. 8 is a flowchart illustrating a method of detecting a target analyte in skin of a subject. The method may include selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to a wavelength absorbed by the target analyte, as in block 810. The target analyte may be at least one of: a vitamin, a mineral, pigment (e.g., melanin), or hemoglobin. Further, the target analyte may be an antioxidant thatThorpe North & Western, LLP Docket No.: 01336-36037.PCT is at least one of: carotenoids, vitamin A, vitamin C, vitamin D, vitamin E, polyphenols, zinc, selenium, oxygen, hemoglobin, proteins (e.g. collagen or elastin), amino acids, or any other agent, compound, molecule, biochemical, structure or material with properties that allow its presence or absence, or amount or concentration, to be measured by quantifying the absorbance or reflectance of light applied thereto.

[0063] Light may be emitted from the one or more LEDs toward a subject’s skin, as in block 820. The one or more LEDs may have a wavelength of between 450 nm and 690 nm. In one example, the one or more LEDs may have 5 LEDs in one embodiment.

[0064] A first light intensify value of the light emitted by the one or more LEDs may be measured with at least one photodiode before the light’s interaction with the subject’s skin, as in block 830. A second light intensify value of the light emitted by the one or more LEDs may be measured with at least one photodiode after the light’s interaction with the subject’s skin, as in block 840. In other words, the light may be measured both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the skin using an emission measuring photodiode, and then 2) as the light returns from the skin using an interaction measuring photodiode.

[0065] A correlation output may be provided for measurements of the light using a machine learning model, 850. The machine learning model may be: regression, decision trees, classification and regression trees (CART), Naive-Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, or supervised machine learning models.

[0066] In another embodiment, light from the one or more LEDs may be passed through a first polarizing filter. Then light that has interacted with the skin may be detected using the at least one photodiode and the light may have a different polarization than the first polarizing filter. For example, the light entering the at least one photodiode may pass through a second polarizing filter with a different spatial orientation or polarizing orientation than the first polarizing filter.

[0067] Each of the one or more LEDs may be activated to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle. Each cycle may be repeated at least 50 times, or at least 100 times, or at least 300 times.

[0068] FIG. 9 is a flowchart illustrating a method for providing a decision output related to tissue of an individual using a spectral light sensor. The method may include providing light from a plurality of light emitting diodes (LEDs) to tissue of the individual, or another sample (such as a calibration standard). The plurality of LEDs may emit light atThorpe North & Western, LLP Docket No.: 01336-36037.PCT wavelengths of 470 nanometers (nm), 495 nm, 520 nm, 590 nm and 630 nm, as in block 910. Thus, each of the plurality of LEDs may emit a different wavelength of light. The tissue may¬ be the skin of a person or animal, or even the tissue of a plant.

[0069] Each of the five wavelengths, or a combination or sub-combination thereof, can be used to test for a different analyte. Each of the five wavelengths, may also be weighted differently when submitted to the machine learning depending on the machine learning model created from the initial training data. The testing for the analyte may be for carotenoids, as described earlier. Further, the testing may also apply for testing to find vitamins, a combination of compounds that might represent a disease (e.g., cancer), or any grouping of analytes that might represent disease in the scanned tissue. This enables a subject to be scanned or tested by the device on a regular basis. The scans can be the basis of predictions for health diagnoses or outcomes defined by7what their tissue score reflected.

[0070] The light emission may include activating each of the plurality' of LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle where only one LED at a time is illuminated. The cycle may be repeated at least 50 times, or at least 100 times, or at least 300 times.

[0071] A plurality of sensor reading values may be obtained for the light from the plurality- of LEDs using a photodiode, including light that has interacted with the tissue of the individual, as in block 920. In one embodiment two photodiodes may be used. Sensor reading values may be obtained from an emission measuring diode with a first spatial orientation to measure light emitted by' the plurality of LEDs and an interaction measuring diode with a second spatial orientation to measure light that has already interacted with tissue.

[0072] The sensor reading values may be processed using a machine learning (ML) model, as in block 930. The machine learning model may be a model that is at least one of: regression, decision trees, classification and regression trees (CART), partial least squares regression, ridge regression, gradient boosted decision trees, Naive-Bayes, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost. or a supervised machine learning.

[0073] A decision output may be received from the machine learning model that represents a health aspect of the tissue, as in block 940. The decision output may represent that the individual has: a defined level of health, a disease, a defined level of a nutrient in the tissue, a vitamin amount in the tissue, a mineral amount in the tissue or an anti-oxidant amount in the tissue.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT

[0074] The ML or artificial intelligence (Al) engine can also be used to compensate for individual variation in tissue. For example, melanin may interfere with the scans. The scan can determine the melanin content using the LED light scanning and adjust the final score (e.g., for carotenoids) accordingly. The wavelength(s) that best detect melanin can be used to detect the melanin in a person's skin for submission to the ML.

[0075] The decision output may be a value representing or otherwise compared to or correlated with a Raman measurement scale. The machine learning may obtain the sensor reading values and the machine learning may be trained using sensor reading values from a plurality of individuals which are each linked with a corresponding Raman measurement output for carotenoids, or other target analyte. The machine learning model can then determine for sensor reading values received for a person what their corresponding Raman measurement is for that person.

[0076] In an example of training the device and the related machine learning models, N test subjects with high carotenoids and M test subject with low carotenoids may be tested or sampled. Each of these subjects (e.g.. people) may be tested by scanning with LED light and photodiode measurements (i.e., reflective spectroscopy) as described herein. These same test subjects or people may also be tested using Raman technology. Thus, person A may receive a first scan with LED light and a first Raman reading, person B may receive a second scan with LED light and a second Raman reading and person C may receive a third scan with LED light and a third Raman reading. This data set is submitted to the machine learning for training of the ML model and the mapping betw een the reflective spectroscopy and the Raman readings. Then the ML generates a model to tie the two data sets together, and the ML determines a specific conversion for reflective spectroscopy measurements to Raman readings.

[0077] As more readings are received for the training data, the Al (e g. Al agent) can be updated over time. The ML model may change based on the training data over a long period of time. The retraining of the ML model should improve the results of the scanning device. This feedback loop may change things such as the weightings used with the light sensor readings (i.e., features) submitted to the ML model or the internal structure of the model used.

[0078] The machine learning model can vary, but in various examples the machine learning model can include a function that outputs the correlation output based on inputs including the photodiode measurements taken using the photodiodes of the analyte measuring device. Any suitable machine learning technique can be used to generate the function. InThorpe North & Western, LLP Docket No.: 01336-36037.PCT various examples, the machine learning model can be trained using a set of data including measurements taken using photodiodes of one or more analyte measuring devices as described herein. The training data set can also include measurements taken using another device, such as a Raman spectroscopy -based scanner. The training data can also include concentrations of target analytes measured in blood serum from one or more subjects. These training data can be used to train the machine learning model, with the goal of generating a function that outputs a correlation output that correlates with the Raman spectroscopy data and / or the blood serum data, based on photodiode measurements taken using an analyte measuring device as described herein.

[0079] Various machine learning techniques can generate functions that can include additive terms having weighting factors. In some examples, the additive terms can include one or more photodiode measurements collected by the analyte measuring device. The weighting factors and arrangement of the terms can be selected when training the machine learning model to produce a best fit of the training data. In certain examples, some terms in the function can include multiple photodiode measurements (e.g.. measurements of multiple different colors of LED) combined in an additive or non-additive way, such as by multiplying two or more photodiode measurements, dividing a photodiode measurement by another measurement, etc. In further examples, the function can include thresholds, decision trees, random forests, gradient boosted trees, rule-based models, or a combination of these.

[0080] The scan readings may be made by measuring reflectance values from the epidermis and / or the dermis. This may include light that reflects from the surface of the skin and light that penetrates the surface of the skin to be scattered in the skin, in which case a portion of the scattered light can be re-emitted back toward the photodiode.

[0081] In another embodiment of the technology, light from the plurality of LEDs can be passed through a first polarizing filter. Unpolarized LED light can include light waves having electric field vectors oriented in random directions. When the light is polarized using a polarizing filter, the electric field vectors of the polarized light are aligned in a single direction. This light may be on the way to interact with (e.g., bounce off of or enter) the tissue. Light that has interacted with the tissue, specifically entered the skin and returned, may be detected using the photodiode and has a different polarization than the first polarizing filter. Furthermore, light entering the photodiode may pass through a second polarizing filter with a different orientation than the first polarizing filter. The second photodiode can look at the light after it passes through a second polarizer. So, the light is reangled by the skin for theThorpe North & Western, LLP Docket No.: 01336-36037.PCT second photodiode to sense the light. That re-polarization only happens if the light goes inside the tissue.

[0082] This technology can also be used to correct for calibration slip. Devices in each country in the world may report back to a cloud-based service on the scans or readings that are being taken. If there are calibration errors between different places in the world, then appropriate adjustments may be made. As an example, if a scanner in Peru is providing average readings at 60K and in Japan the average reading is 55K. Then a ML model can receive the readings from around the world and calibrate the devices globally. This may mean the average reading in Japan can be adjusted by 5K to calibrate back to scans from the entire world.

[0083] FIG. 10 is flowchart illustrating an example of a method for scanning tissue of a subject. Light may be emitted from a plurality of light emitting diodes (LEDs) toward a subject’s tissue, as in block 1010. The plurality of light emitting diodes (LEDs) sends a sequence of separate light wavelengths to interact with the subject’s tissue.

[0084] Measurements from the light emitted from the LEDs may be detected with at least one photodiode and the light returning from interacting with the subject’s tissue, as in block 1020. The light may be measured at two points in the lights path. The first measurement may be made as the light from the plurality of light emitting diodes (LEDs) is directed toward the subject’s tissue using an emission measuring photodiode but before the light reaches the subject’s tissue. The second measurement may be made as the light returns from the subject’s tissue using an interaction measuring photodiode.

[0085] A correlation output may be provided for the measurements of the light using a machine learning model, as in block 1030. The correlation output may be: a value representing antioxidants in the subject’s tissue, a value representing vitamins in the subject’s tissue, a value representing minerals in the subject’s tissue, a tissue condition, or a disease. The machine learning model may be models like: regression, decision trees, classification and regression trees (CART), Naive-Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, or supervised machine learning.

[0086] FIG. 11 is a flowchart illustrating an example of a method of training a machine learning model to provide a prediction for a target analyte in an area of a subject’s skin. The method may include obtaining a plurality of sensor readings from a plurality of subjects using a reflective spectroscopy device generating one or more wavelengths of light corresponding to a wav elength absorbed by the target analyte, as in block 1110.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT

[0087] A plurality of reference measurements may be obtained for the analyte, as in block 1120. The reference measurements may provide an accurate portrayal of the analyte in the person’s skin using another measurement process or device. The reference measurements can be associated with the sensor measurements, and the sensor measurements may be a ratio of a first light intensity value and a second light intensity value that is measured using photodiodes. Each sensor measurement (e.g., ratio or other calculation) may be associated with a reference measurement and may be provided as a training sample or group of features to train the machine learning model. Hundreds, thousands or millions of samples may be used for completely training the machine learning model. The reference measurements may be measurements from a blood sample, a Raman spectrometer or another trusted measurement for the analyte. It is desirable for the reference measurements to provide an accurate value of the analyte being measured, whether that is vitamin C, vitamin D, minerals, carotenoids, or other analytes. As a result, the reference measurements can be used to train the machine learning model.

[0088] A machine learning model may be trained using the plurality of sensor readings and the reference measurements to enable the machine learning model to provide a prediction output, as in block 1130. The sensor readings (e.g., a measurement ratio or another calculation) and reference measurements are features of each training sample and are submitted to the machine learning model to train the machine learning model to output reference measurements or labels when new sensor readings are received for a subject’s skin. As discussed earlier, detecting the first light intensity values and the second light intensity values may occur thousands of times. These values may be combined (e.g., averaged) and then calculated into a ratio.

[0089] The operation of obtaining a plurality of sensor readings may include further actions. For example, one or more light emitting diodes (LEDs) may be selected having one or more wavelengths corresponding to the wavelength absorbed by the target analyte. Light may be emitted from the one or more LEDs toward a plurality of subjects’ skin. At least one photodiode may be used to measure first light intensity values of the light emitted by the one or more LEDs before the light’s interaction with the plurality of subject’s skins. Another photodiode may measure second light intensity values of the light emitted by the one or more LEDs after the light’s interaction with the plurality of subjects’ skins to provide a plurality of sensor readings.

[0090] In some examples, the sensor readings may the represented as ratios. More specifically, the spectroscopy device may compute or provide ratio outputs based on firstThorpe North & Western, LLP Docket No.: 01336-36037.PCT light intensity values and second light intensity values detected for each of the plurality of subjects.

[0091] The prediction output of the machine learning model may be a reference measurement, a label or an abstract representation. In order to use the trained machine learning model, a sensor reading may be obtained for a subject's skin. Then a prediction output may be provided for the sensor reading for the subject’s skin using the machine learning model.

[0092] After a machine learning model has been trained using data such as photodiode measurements collected using the devices described herein, which may be input into the machine learning model together with comparison values collected using Raman spectroscopy and / or analyte concentrations measured by blood tests and / or other data, the machine learning model can then be used by the analyte measuring device to calculate a correlation output as described above. In some examples, the machine learning model can be programmed into the analyte measuring device so that the analyte measuring device calculates the correlation output based on photodiode measurements taken during a scan of a subject’s skin. The machine learning model can be a function into which the photodiode data are input as parameters, and the function can output the correlation output. As explained above, the correlation output can include a numeric score, an alpha-numeric score, a classification, or a combination thereof. It is noted that in some examples the analyte measuring device can utilize the machine learning model without further updating or training the machine learning model. At this point, the machine learning model can be static, meaning that no further training of the model is performed. How ever, in other examples, the photodiode measurements taken using the analyte measuring device can also be used as training data to further train the machine learning model. In certain examples, the photodiode measurement data can be uploaded from the analyte measurement device to another computing device such as a personal computer, cloud computing system, server, or other computing device and the further training of the machine learning model can be performed by this other computing device.

[0093] In a particular example, the correlation output can represent a carotenoid signal estimation, or in other words, an estimated score correlating to the level carotenoids in the skin. This can be estimated using a computational model comprising multivariate machine learning and artificial intelligence based regression and classification methods trained on curated datasets. The model can be configured to identify and quantify the carotenoid associated absorption signature from the multispectral measurements taken usingThorpe North & Western, LLP Docket No.: 01336-36037.PCT the analyte measuring device while explicitly representing, disentangling, and subtracting confounding contributions from non-carotenoid chromophores, for example hemoglobin and other optically active constituents, and other nuisance terms such as baseline scattering and pathlength variability. In operation, the model can process band specific measurements using a dark subtracted ratiometric metric defined as the reflected signal divided by the corresponding monitored LED radiant output for each of the color bands of the LEDs. In certain examples, four different colored LEDs can be used. Band-dependent constants are applied to the resulting ratios to yield four fixed hard-coded parameters. These parameters can be identical across all units of the analyte measuring device. In addition, the computation can incorporate a single unit specific calibration parameter that is determined for each individual device and applied to the final estimation pathway to enhance inter-unit agreement, reduce systematic bias, and improve precision across production variance.

[0094] Some of the functional units described in this specification have been labeled as modules, in order to more particularly emphasize their implementation independence. For example, a module may be implemented as a hardware circuit comprising custom VLSI circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors, or other discrete components. A module may also be implemented in programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices or the like.

[0095] Modules may also be implemented in software for execution by various types of processors. An identified module of executable code may, for instance, comprise one or more blocks of computer instructions, which may be organized as an object, procedure, or function. Nevertheless, the executables of an identified module need not be physically located together, but may comprise disparate instructions stored in different locations which comprise the module and achieve the stated purpose for the module when joined logically together.

[0096] Indeed, a module of executable code may be a single instruction, or many instructions, and may even be distributed over several different code segments, among different programs, and across several memory devices. Similarly, operational data may be identified and illustrated herein within modules, and may be embodied in any suitable form and organized within any suitable type of data structure. The operational data may be collected as a single data set or may be distributed over different locations including over different storage devices. The modules may be passive or active, including agents operable to perform desired functions.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT

[0097] The technology described here can also be stored on a computer readable storage medium that includes volatile and non-volatile, removable and non-removable media implemented with any technology for the storage of information such as computer readable instructions, data structures, program modules, or other data. Computer readable storage media include, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM. digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other computer storage medium which can be used to store the desired information and described technology.

[0098] The devices described herein may also contain communication connections or networking apparatus and networking connections that allow the devices to communicate with other devices. Communication connections are an example of communication media. Communication media typically embodies computer readable instructions, data structures, program modules and other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. A “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired netw ork or direct-wired connection, and wireless media such as acoustic, radio frequency, infrared, and other wireless media. The term computer readable media as used herein includes communication media.

[0099] Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more examples. In the preceding description, numerous specific details were provided, such as examples of various configurations to provide a thorough understanding of examples of the described technology. One skilled in the relevant art will recognize, however, that the technology can be practiced without one or more of the specific details, or with other methods, components, devices, etc. In other instances, well-known structures or operations are not shown or described in detail to avoid obscuring aspects of the technology.

[0100] Although the subject matter has been described in language specific to structural features and / or operations, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features and operations described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims. Numerous modifications and alternativeThorpe North & Western, LLP Docket No.: 01336-36037.PCT arrangements can be devised without departing from the spirit and scope of the described technology.Examples

[0101] Example 1: A study was performed to validate an example analyte measurement device as described herein. The study included 46 participants. Of these, 29 participants received a supplement (LifePak) containing vitamins, minerals, carotenoids, and phytonutrients, and 17 participants received a placebo. The subjects were instructed to take the supplement or placebo two times daily with meals for 12 weeks. All subjects were scanned using the example analyte measurement device (referred to as the Prysm scanner in this study) and with a Pharmanex BioPhotonic S3 scanner (Raman Spectrometry Scanner), which is another scanner based on Raman spectroscopy. At week 12, all subjects had blood draw n to measure levels of vitamin C and selenium. Subjects filled out questionnaires about quality of life after w eeks 1, 2, 4, 8, and 12.

[0102] Both the Prysm scanner and the S3 scanner showed similar patterns in tracking changes in skin carotenoid levels over time. The supplement group showed significant improvement in carotenoid status as indicated by an increase in the S3 scanner scores and Prysm scanner scores at week 12. Table 1 shows the average scores of the supplement group and the placebo group measured with the S3 scanner at weeks 0, 1, 2. 4, 8, and 12. Table 2 shows the same scores as measured with the Prysm scanner.Table 1: S3 ScannerThorpe North & Western, LLP Docket No.: 01336-36037.PCT Table 2; Prysm Scanner

[0103] The supplement group had a 44% increase in serum vitamin C levels and 26% increase in serum selenium levels with no change in the placebo group for either nutrient. At week 12. a high percent of the subjects in the supplement group reported quality of life improvements. These results suggest that both the S3 scanner and the Prysm scanner were able to detect changes in skin carotenoid level due to taking the supplement. The supplement appears to have been well-absorbed and had a meaningful impact on nutrient status and quality of life.

[0104] Example 2: Another study was performed with 83 subjects between the ages of 3 and 17. Each subject was scanned two times with the Pharmanex BioPhotonic S3 scanner and then scanned two times with an analyte measuring device as described herein (referred to in this study as a Prysm scanner). The two scans with each scanner include one scan of the index finger and one scan of the thumb. FIG. 12A is a graph of S3 scores vs. Prysm scores for scans of the index finger. The S3 scores correlated with the Prysm scores with a R2value of 0.57. FIG. 12B is a graph of S3 scores vs. Prysm scores for scans of the thumb. The S3 scores correlated with the Prysm scores with an R2value of 0.6. These results show a strong correlation between the scores obtained with the S3 scanner and the Prysm scanner. This suggests that the analyte measuring devices described herein can be used to accurately measure analyte levels in the skin comparable to the Raman spectroscopy -based S3 scanner.

[0105] Example 3: A dose-response study was performed. Subjects were scanned using a Pharmanex BioPhotonic S3 scanner and an analyte measuring device as described herein (referred to in this study as a Prysm scanner). Scans were performed at weeks 0, 1, 2, 4, 8, and 12. Subjects also filled out a questionnaire at the same time as each scan. A controlThorpe North & Western, LLP Docket No.: 01336-36037.PCT group (Group 1) included 20 subjects that did not take any supplement. Group 2 included 45 subjects that took one antioxidant strip daily. Group 3 included 45 subjects that took two antioxidant strips daily. Group 4 included 45 subjects that took three antioxidant strips daily. The antioxidant strips contained lycopene, lutein, and beta carotene. FIG. 13 A shows the scores measured with the Prysm scanner and the S3 scanner over time for the control group. FIG. 13B shows the scores for Group 2, FIG. 13C shows the scores for Group 3, and FIG. 13D shows the scores for Group 4. On average across all the measurements, the Prysm scanner scores correlated to the S3 scanner scores with a R2value of 0.75. These results suggest that the supplement antioxidant strips increased carotenoid levels in the subjects and that the Prysm scanner is comparable to the S3 scanner for measuring carotenoid levels.

[0106] Example 4: A new machine learning model was trained using previously collected data. The machine learning model was programmed into two analyte measuring devices as described herein (referred to in this example as Prysm scanners). The Prysm scanners were then used to scan the skin of a group of subjects and generate a score representing the carotenoid level in the skin. Blood samples were taken from the same group of subjects and the serum carotenoid concentration was measured. FIG. 14A shows the serum carotenoid concentrations vs. Prysm scores for the first scanner. These values correlate with a R2value of 0.7659. FIG. 14B shows the serum carotenoid concentrations vs. Prysm scores for the second scanner. These values correlate with a R2value of 0.7543. These results suggest that the two scanners provide similar scan results. The machine learning model provided a strong correlation to the serum carotenoid levels. The Prysm scores calculated using the second scanner were also compared to S3 scanner scores measured for the same subjects. FIG. 14C shows a graph of the Prysm scanner scores vs. the S3 scanner scores. These scores correlated with a R2value of 0.8404.

[0107] Additional Examples: The technology described herein can include any of the examples in the following numbered list:

[0108] Example 1. A method of detecting a target analyte in a subject’s skin, comprising:selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to a wavelength absorbed by the target analyte;emitting light from the one or more LEDs toward the subject’s skin;measuring with at least one photodiode a first light intensity value of the light emitted by the one or more LEDs before the light’s interaction with the subject’s skin;Thorpe North & Western, LLP Docket No.: 01336-36037.PCT measuring with at least one photodiode a second light intensity value of the light emitted by the one or more LEDs after the light’s interaction with the subject's skin; and providing a correlation output for measurements of the light using a machine learning model.Example 2. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the target analyte is at least one of: a vitamin, a mineral, a pigment, or hemoglobin.Example 3. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the target analyte is an antioxidant that is at least one of: carotenoids, vitamin A, vitamin C, vitamin D, vitamin E. polyphenols, zinc, or selenium.Example 4. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the one or more LEDs have a wavelength of between 450 nm and 690 nm.Example 5. The method as in any of examples 1-44 or 61. or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the one or more LEDs is an array of LEDs.Example 6. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the one or more LEDs has 5 LEDs.Example 7. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising passing light from the one or more LEDs through a first polarizing filter.Example 8. The method as in any of examples 1-44 or 61. or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising detecting light that has interacted with the subject’s skin and passed through a second polarizing filter prior to the second photodiode measurement.Example 9. The method as in any of examples 1-44 or 61. or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising measuring light both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the subject's skin using an emission measuring photodiode, and 2) as the light returns from the subject's skin using an interaction measuring photodiode.Example 10. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising activatingThorpe North & Western, LLP Docket No.: 01336-36037.PCT each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.Example 11. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the cycle is repeated at least 50 times.Example 12. A method, comprising:emitting light from one or more light emitting diodes (LEDs) toward a subject’s tissue;detecting with at least one photodiode, measurements from the light emitted from the one or more LEDs, including the light returning from interacting with the subject’s tissue; andproviding a correlation output for the measurements of the light using a machine learning model.Example 13. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the correlation output is at least one of: a value representing antioxidants in the subject’s tissue, a value representing vitamins in the subject’s tissue, a value representing minerals in the subject’s tissue, a tissue condition, or a disease.Example 14. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising using the machine learning model that is at least one of: regression, decision trees, classification and regression trees (CART), Naive-Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, supervised machine learning.Example 15. The method as in any of examples 1 -44 or 61 , or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the one or more LEDs is a plurality of LEDs that sends a sequence of separate light wavelengths to interact with the subject's tissue.Example 16. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising measuring light both: 1) from the one or more LEDs as the light is directed toward the subject’s tissue using an emission measuring photodiode, and 2) as the light returns from the subject’s tissue using an interaction measuring photodiode.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT Example 17. A method for providing a decision output related to tissue of an individual using a spectral light sensor, comprising:providing light from one or more light emitting diodes (LEDs) to tissue of the individual, wherein the one or more LEDs emit light at wavelengths of one or more of 470 nanometers (nm), 495 nm, 520 nm, 590 nm and 630 nm;obtaining a plurality of sensor reading values using a photodiode for the light from the one or more LEDs, including light that has interacted with the tissue of the individual;processing the sensor reading values using a machine learning model; and receiving the decision output from the machine learning model that represents a health aspect of the tissue.Example 18. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein obtaining the sensor reading values using the photodiode, further comprises obtaining sensor reading values from an emission measuring diode with a first spatial orientation to measure light emitted by the one or more LEDs and an interaction measuring diode with a second spatial orientation to measure light that has interacted with tissue.Example 19. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein providing light further comprises activating each of the one or more LEDs to provide one or more LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.Example 20. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the cycle is repeated at least 50 times.Example 21. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein each of the one or more LEDs turns on and off sequentially to form a cycle.Example 22. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein only one LED at a time is illuminated.Example 23. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the one or more LEDs each emit a different w avelength of light.Example 24. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the decision outputThorpe North & Western, LLP Docket No.: 01336-36037.PCT represents that the individual has at least one of: a defined level of health, a disease, a nutrient in the tissue, a vitamin amount in the tissue, or a mineral amount in the tissue.Example 25. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the decision output is a value representing a Raman measurement scale.Example 26. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the machine learning model is trained using optical readings from a plurality of individuals which are linked with a corresponding Raman measurement scale for carotenoids.Example 27. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising using the machine learning model that is at least one of: regression, decision trees, classification and regression trees (CART), Naive-Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, supervised machine learning.Example 28. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the tissue is skin of a person or animal or tissue of a plant.Example 29. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising passing light from the one or more LEDs through a first polarizing filter.Example 30. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising detecting light that has interacted with the tissue using the photodiode and has a different polarization than the first polarizing filter.Example 31. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein light entering the photodiode passes through a second polarizing filter with a different orientation than the first polarizing filter.Example 32. A method of training a machine learning model to provide a prediction for a target analyte in an area of a subject’s skin, comprising:obtaining a plurality of sensor readings from a plurality of subjects using a reflective spectroscopy device generating one or more wavelengths of light corresponding to a wavelength absorbed by the target analyte;Thorpe North & Western, LLP Docket No.: 01336-36037.PCT obtaining a plurality of reference measurements for the target analyte; and training a machine learning model using the plurality of sensor readings and the reference measurements to enable the machine learning model to provide a prediction output.Example 33. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein obtaining a plurality' of sensor readings further comprises:selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to the wavelength absorbed by the target analyte;emitting light from the one or more LEDs toward a plurality of subjects’ skin; measuring with at least one photodiode first light intensity values of the light emitted by the one or more LEDs before the light’s interaction with the plurality of subject’s skins; andmeasuring with at least one photodiode second light intensity values of the light emitted by the one or more LEDs after the light’s interaction with the plurality of subjects' skins to provide a plurality of sensor readings.Example 34. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the reflective spectroscopy device provides ratio outputs based on first light intensity values and second light intensity’ values detected for each of the plurality of subjects.Example 35. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, yvherein the prediction output is a reference measurement or a label.Example 36. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. further comprising:obtaining a sensor reading for a subject’s skin; andproviding a prediction output for the sensor reading for the subject’s skin using the machine learning model.Example 37. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the target analyte is at least one of: a vitamin, a mineral, melanin, or hemoglobin.Example 38. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the target analyte is an antioxidant that is at least one of: carotenoids, vitamin A, vitamin C, vitamin D, vitamin E. polyphenols, zinc, or selenium.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT Example 39. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising passing light from the one or more LEDs through a first polarizing filter.Example 40. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising detecting light that has interacted with the subject’s skin using a photodiode and has a different polarization than the first polarizing filter.Example 41. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein light entering the photodiode passes through a second polarizing filter with a different spatial orientation than the first polarizing filter.Example 42. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising measuring light both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the subject’s skin using an emission measuring photodiode, and 2) as the light returns from the subject’s skin using an interaction measuring photodiode.Example 43. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising activating each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.Example 44. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the cycle is repeated at least 50 times.Example 45. An analyte measuring device, comprising:a housing having an aperture;one or more LEDs within the housing oriented to emit light through the aperture toward a subject's skin, wherein the LEDs have one or more wavelengths corresponding to wavelengths absorbed by a target analyte in the subject’s skin;a first photodiode within the housing oriented to measure a first light intensity of the light emitted by the one or more LEDs before the light interacts with the subject’s skin; a second photodiode within the housing oriented to measure a second light intensity7value of the light emitted by the one or more LEDs after the light interacts with the subject’s skin; andThorpe North & Western, LLP Docket No.: 01336-36037.PCT a processor in electronic communication with the first photodiode and the second photodiode, wherein the processor is programmed to provide a correlation output for the measurements by the first photodiode and the second photodiode using a machine learning model.Example 46. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the correlation output is at least one of: a value representing antioxidants in the subject’s tissue, a value representing vitamins in the subject’s tissue, a value representing minerals in the subject’s tissue, a tissue condition, or a disease.Example 47. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. wherein the one or more LEDs have a wavelength of between 450 nm and 690 nm.Example 48. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60. further comprising a first polarizing filter positioned to polarize the light emitted by the one or more LEDs before the light interacts with the subject’s skin.Example 49. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising a second polarizing filter positioned to polarize the light emitted by the one or more LEDs after the light interacts with the subject’s skin before being received by the second photodiode.Example 50. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the first polarizing filter and the second polarizing filter are cross polarizing with respect to each other.Example 51. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising a separation panel between the one or more LEDs and the second photodiode to prevent direct exposure of the second photodiode to light emitted by the one or more LEDs.Example 52. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the processor is further programmed to activate each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.Example 53. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the LED color emission states comprise separate light wavelengths emitted by individual LEDs, and wherein theThorpe North & Western, LLP Docket No.: 01336-36037.PCT processor is further programmed to measure an intensity of each separate light wavelength using the first photodiode and the second photodiode.Example 54. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the correlation output comprises a numeric value.Example 55. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising a display electronically connected to the processor, wherein the processor is further programmed to display the numeric value on the display.Example 56. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising an array of indicator LEDs electronically connected to the processor, wherein the processor is further programmed to activate the indicator LEDs to indicate scan progress, the correlation output, or a combination thereof.Example 57. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, further comprising a wireless communication module electronically connected to the processor, wherein the processor is further programmed to transmit the correlation output to an external computing device using the wireless communication module.Example 58. An analyte measuring system, comprising:a skin scanning device comprising:a housing having an aperture,one or more LEDs within the housing oriented to emit light through the aperture toward a subject’s skin, wherein the LEDs have one or more wavelengths corresponding to wavelengths absorbed by a target analyte in the subject’s skin,a first photodiode within the housing oriented to measure a first light intensity of the light emitted by the one or more LEDs before the light interacts with the subject’s skin,a second photodiode within the housing oriented to measure a second light intensity7value of the light emitted by the one or more LEDs after the light interacts with the subject’s skin; anda correlation module in electronic communication with the skin scanning device, wherein the correlation module is programmed to provide a correlation output for theThorpe North & Western, LLP Docket No.: 01336-36037.PCT measurements by the first photodiode and the second photodiode using a machine learning model.Example 59. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the correlation module comprises a processor integrated in the skin scanning device, a mobile device app, a web app, a computer application, or a combination thereof.Example 60. The method as in any of examples 1-44 or 61, or the device as in any of examples 45-57, or the system as in any of examples 58-60, wherein the correlation module is further programmed to update the machine learning model using the measurements by the first photodiode and the second photodiode.Example 61, A method of increasing a level of a target analyte in a subject’s skin using the system of claim 58, the method comprising:scanning the subject’s skin using the skin scanning device;providing the correlation output using the correlation module;identifying a deficiency of the target analyte based on the correlation output; and supplying the subject with a supplement comprising the target analyte.

[0109] While the forgoing examples are illustrative of the principles of the present technology in one or more particular applications, it will be apparent to those of ordinary skill in the art that numerous modifications in form, usage and details of implementation can be made without the exercise of inventive faculty7, and without departing from the principles and concepts of the technology.

Claims

1. Thorpe North & Western, LLP Docket No.: 01336-36037.PCT CLAIMSWhat is claimed is:

1. A method of detecting a target analyte in a subject’s skin, comprising:selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to a wavelength absorbed by the target analyte;emitting light from the one or more LEDs toward the subject’s skin; measuring with at least one photodiode a first light intensity value of the light emitted by the one or more LEDs before the light's interaction with the subject’s skin; measuring with at least one photodiode a second light intensity value of the light emitted by the one or more LEDs after the light’s interaction with the subject’s skin; and providing a correlation output for measurements of the light using a machine learning model.

2. The method as in claim 1, wherein the target analyte is at least one of: a vitamin, a mineral, a pigment, or hemoglobin.

3. The method as in claim 1, wherein the target analyte is an antioxidant that is at least one of: carotenoids, vitamin A. vitamin C, vitamin D, vitamin E, polyphenols, zinc, or selenium.

4. The method as in claim 1, wherein the one or more LEDs have a wavelength of between 450 nm and 690 nm.

5. The method as in claim 1, wherein the one or more LEDs is an array of LEDs.

6. The method of claim 5, wherein the one or more LEDs has 5 LEDs.

7. The method of claim 1, wherein at least one of the one or more LEDs emits multiple different colors.

8. The method as in claim 1, further comprising passing light from the one or more LEDs through a first polarizing filter.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT 9. The method as in claim 8, further comprising detecting light that has interacted with the subject's skin and passed through a second polarizing fdter prior to the second photodiode measurement.

10. The method as in claim 1, further comprising measuring light both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the subject’s skin using an emission measuring photodiode, and 2) as the light returns from the subject’s skin using an interaction measuring photodiode.

11. The method as in claim 1, further comprising activating each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.

12. The method as in claim 11, wherein the cycle is repeated at least 50 times.

13. A method, comprising:emitting light from one or more light emitting diodes (LEDs) toward a subject’s tissue;detecting with at least one photodiode, measurements from the light emitted from the one or more LEDs, including the light returning from interacting with the subject’s tissue; andproviding a correlation output for the measurements of the light using a machine learning model.

14. The method as in claim 13, wherein the correlation output is at least one of: a value representing antioxidants in the subject’s tissue, a value representing vitamins in the subject’s tissue, a value representing minerals in the subject’s tissue, a tissue condition, or a disease.

15. The method as in claim 13, further comprising using the machine learning model that is at least one of: regression, decision trees, classification and regression trees (CART), Naive- Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, supervised machine learning.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT16. The method as in claim 13, wherein the one or more LEDs is a plurality of LEDs that sends a sequence of separate light wavelengths to interact with the subject’s tissue.

17. The method as in claim 13, further comprising measuring light both: 1) from the one or more LEDs as the light is directed toward the subject’s tissue using an emission measuring photodiode, and 2) as the light returns from the subject’s tissue using an interaction measuring photodiode.

18. A method for providing a decision output related to tissue of an individual using a spectral light sensor, comprising:providing light from one or more light emitting diodes (LEDs) to tissue of the individual, wherein the one or more LEDs emit light at wavelengths of one or more of 470 nanometers (nm), 495 nm, 520 nm, 590 nm and 630 nm;obtaining a plurality of sensor reading values using a photodiode for the light from the one or more LEDs, including light that has interacted with the tissue of the individual;processing the sensor reading values using a machine learning model; and receiving the decision output from the machine learning model that represents a health aspect of the tissue.

19. The method as in claim 18, wherein obtaining the sensor reading values using the photodiode, further comprises obtaining sensor reading values from an emission measuring diode with a first spatial orientation to measure light emitted by the one or more LEDs and an interaction measuring diode with a second spatial orientation to measure light that has interacted with tissue.

20. The method as in claim 18, wherein providing light further comprises activating each of the one or more LEDs to provide one or more LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.

21. The method as in claim 20, wherein the cycle is repeated at least 50 times.

22. The method as in claim 20, wherein each of the one or more LEDs turns on and off sequentially to form a cycle.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT 23. The method as in claim 22, wherein only one LED at a time is illuminated.

24. The method as in claim 18, wherein the one or more LEDs each emit a different wavelength of light.

25. The method as in claim 18, wherein the decision output represents that the individual has at least one of: a defined level of health, a disease, a nutrient in the tissue, a vitamin amount in the tissue, or a mineral amount in the tissue.

26. The method of claim 18, wherein the decision output is a value representing a Raman measurement scale.

27. The method as in claim 18, wherein the machine learning model is trained using optical readings from a plurality of individuals which are linked with a corresponding Raman measurement scale for carotenoids.

28. The method as in claim 18, further comprising using the machine learning model that is at least one of: regression, decision trees, classification and regression trees (CART), Naive- Bay es, K-nearest neighbors, support vector machines (SVG), a neural network, classification models, ensemble models, random tree forests, Boosting with XGBoost, supervised machine learning.

29. The method as in claim 18, wherein the tissue is skin of a person or animal or tissue of a plant.

30. The method as in claim 17, further comprising passing light from the one or more LEDs through a first polarizing filter.

31. The method as in claim 30, further comprising detecting light that has interacted with the tissue using the photodiode and has a different polarization than the first polarizing filter.

32. The method as in claim 30, wherein light entering the photodiode passes through a second polarizing filter with a different orientation than the first polarizing filter.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT 33. A method of training a machine learning model to provide a prediction for a target analyte in an area of a subject’s skin, comprising:obtaining a plurality of sensor readings from a plurality of subjects using a reflective spectroscopy device generating one or more wavelengths of light corresponding to a wavelength absorbed by the target analyte;obtaining a plurality of reference measurements for the target analyte: and training a machine learning model using the plurality of sensor readings and the reference measurements to enable the machine learning model to provide a prediction output.

34. The method as in claim 33, wherein the prediction output is a reference measurement or a label.

35. The method as in claim 33, further comprising:obtaining a sensor reading for a subject’s skin: andproviding a prediction output for the sensor reading for the subject’s skin using the machine learning model.

36. The method as in claim 33, wherein the target analyte is at least one of: a vitamin, a mineral, melanin, or hemoglobin.

37. The method as in claim 33, wherein the target analyte is an antioxidant that is at least one of: carotenoids, vitamin A, vitamin C, vitamin D, vitamin E, polyphenols, zinc, or selenium.

38. The method as in claim 33, wherein obtaining a plurality7of sensor readings further comprises:selecting one or more light emitting diodes (LEDs) having one or more wavelengths corresponding to the wavelength absorbed by the target analyte;emitting light from the one or more LEDs toward a plurality7of subjects’ skin; measuring with at least one photodiode first light intensity7values of the light emitted by the one or more LEDs before the light's interaction with the plurality of subject’s skins; andThorpe North & Western, LLP Docket No.: 01336-36037.PCT measuring with at least one photodiode second light intensity values of the light emitted by the one or more LEDs after the light’s interaction with the plurality of subjects’ skins to provide a plurality of sensor readings.

39. The method as in claim 38, wherein the reflective spectroscopy device provides ratio outputs based on first light intensity values and second light intensity values detected for each of the plurality of subjects.

40. The method as in claim 38, further comprising passing light from the one or more LEDs through a first polarizing filter.

41. The method as in claim 40, further comprising detecting light that has interacted with the subject's skin using a photodiode and has a different polarization than the first polarizing filter.

42. The method as in claim 41, wherein light entering the photodiode passes through a second polarizing filter with a different spatial orientation than the first polarizing filter.

43. The method as in claim 38, further comprising measuring light both: 1) from the one or more light emitting diodes (LEDs) as the light is directed toward the subject’s skin using an emission measuring photodiode, and 2) as the light returns from the subject’s skin using an interaction measuring photodiode.

44. The method as in claim 38, further comprising activating each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.

45. The method as in claim 44, wherein the cycle is repeated at least 50 times.

46. An analyte measuring device, comprising:a housing having an aperture;one or more LEDs within the housing oriented to emit light through the aperture toward a subject’s skin, wherein the LEDs have one or more wavelengths corresponding to wavelengths absorbed by a target analyte in the subject’s skin;Thorpe North & Western, LLP Docket No.: 01336-36037.PCT a first photodiode within the housing oriented to measure a first light intensity of the light emitted by the one or more LEDs before the light interacts with the subject’s skin;a second photodiode within the housing oriented to measure a second light intensity value of the light emitted by the one or more LEDs after the light interacts with the subject’s skin; anda processor in electronic communication with the first photodiode and the second photodiode, wherein the processor is programmed to provide a correlation output for the measurements by the first photodiode and the second photodiode using a machine learning model.

47. The device of claim 46, wherein the correlation output is at least one of: a value representing antioxidants in the subject’s tissue, a value representing vitamins in the subject’s tissue, a value representing minerals in the subject’s tissue, a tissue condition, or a disease.

48. The device of claim 46, wherein the one or more LEDs have a wavelength of between 450 nm and 690 nm.

49. The device of claim 46, further comprising a first polarizing filter positioned to polarize the light emitted by the one or more LEDs before the light interacts with the subject’s skin.

50. The device of claim 49, further comprising a second polarizing filter positioned to polarize the light emitted by the one or more LEDs after the light interacts with the subject’s skin before being received by the second photodiode.

51. The device of claim 50, wherein the first polarizing filter and the second polarizing filter are cross polarizing with respect to each other.

52. The device of claim 46, further comprising a separation panel between the one or more LEDs and the second photodiode to prevent direct exposure of the second photodiode to light emitted by the one or more LEDs.Thorpe North & Western, LLP Docket No.: 01336-36037.PCT 53. The device of claim 46, wherein the processor is further programmed to activate each of the one or more LEDs to provide a plurality of LED color emission states and an off-state lasting multiple milliseconds (ms) each to form a cycle.

54. The device of claim 53, wherein the LED color emission states comprise separate light wavelengths emitted by individual LEDs, and wherein the processor is further programmed to measure an intensity of each separate light wavelength using the first photodiode and the second photodiode.

55. The device of claim 46, wherein the correlation output comprises a numeric value.

56. The device of claim 55, further comprising a display electronically connected to the processor, wherein the processor is further programmed to display the numeric value on the display.

57. The device of claim 46, further comprising an array of indicator LEDs electronically connected to the processor, wherein the processor is further programmed to activate the indicator LEDs to indicate scan progress, the correlation output, or a combination thereof.

58. The device of claim 46, further comprising a wireless communication module electronically connected to the processor, wherein the processor is further programmed to transmit the correlation output to an external computing device using the wireless communication module.

59. An analyte measuring system, comprising:a skin scanning device comprising:a housing having an aperture,one or more LEDs within the housing oriented to emit light through the aperture toward a subject’s skin, wherein the LEDs have one or more wavelengths corresponding to wavelengths absorbed by a target analyte in the subject’s skin,a first photodiode within the housing oriented to measure a first light intensity of the light emitted by the one or more LEDs before the light interacts with the subject’s skin,Thorpe North & Western, LLP Docket No.: 01336-36037.PCT a second photodiode within the housing oriented to measure a second light intensity value of the light emitted by the one or more LEDs after the light interacts with the subject’s skin; anda correlation module in electronic communication with the skin scanning device, wherein the correlation module is programmed to provide a correlation output for the measurements by the first photodiode and the second photodiode using a machine learning model.

60. The system of claim 59, wherein the correlation module comprises a processor integrated in the skin scanning device, a mobile device app, a web app, a computer application, or a combination thereof.

61. The system of claim 59, wherein the correlation module is further programmed to update the machine learning model using the measurements by the first photodiode and the second photodiode.

62. A method of increasing a level of a target analyte in a subject’s skin using the system of claim 59, the method comprising:scanning the subject’s skin using the skin scanning device;providing the correlation output using the correlation module;identifying a deficiency of the target analyte based on the correlation output; and supplying the subject with a supplement comprising the target analyte.

Images