Infrared based core body temperature sensing system and method

Abstract

A non-invasive and a non-contact core body temperature monitoring method and system including an infrared camera, a processor, and a non-transitory computer readable medium with a program for analyzing thermal images of the human eye. The method for using the system that includes isolating ocular thermal images from a facial image; extracting thermal information about the surface of the eye; transitioning the thermal information to a selected model of the eye; estimating a temperature at the back of the eye; optimizing the estimated temperature at the back of the eye, and transmitting the optimized temperature information to an output device.

Claims

1. A system for measuring a body temperature of a subject based on thermal data collected by an infrared camera, the system comprising: a processing unit comprising one or more processors; one or more memories storing at least one three-dimensional model representing a geometry of an eye and instructions configured to be executed by the processing unit, the three-dimensional model comprising a model ocular surface and a model back portion opposite the model ocular surface, wherein the system is configured to: receive the thermal data collected by the infrared camera, the thermal data comprising a plurality of thermal measurements having respective pixel positions; determine an ocular surface portion of the thermal data that corresponds to a subject ocular surface of the subject; transition, based on the pixel positions, the thermal measurements of the ocular surface portion of the thermal data to locations on the model ocular surface on the three-dimensional model representing the geometry of the eye that correspond to locations on the subject ocular surface from which the thermal measurements were collected; and based on the thermal data, as transitioned to the model ocular surface of the three-dimensional model, determine a temperature at the model back portion of the three-dimensional model, wherein the determined temperature at the model back portion indicates the body temperature of the subject; and transmit the determined temperature to an output device.

2. The system of claim 1, wherein determining the temperature at the model back portion of the three-dimensional model comprises calculating what the temperature at the model back portion of the three-dimensional model must be to produce the observed thermal data, as transitioned to the model ocular surface, based on known boundary conditions and environmental conditions.

3. The system of claim 2, wherein the environmental conditions include a convective heat transfer parameter of air adjacent to the subject.

4. The system of claim 1, wherein the system is further configured to obtain the thermal data using the infrared camera.

5. The system of claim 1, wherein determining an ocular surface portion of the thermal data is performed using an optical image of the subject that indicates a location of the subject ocular surface of the subject.

6. The system of claim 5, wherein the system is further configured to process the thermal data and/or the optical image to correct for differences in one or more of: (i) image sizes, (ii) orientations, or (iii) fields of view.

7. The system of claim 5, wherein the system is further configured to obtain the thermal data using the infrared camera and to obtain the optical image using an optical camera.

8. The system of claim 1, wherein the determined temperature indicates a core body temperature of the subject.

9. The system of claim 1, wherein the three-dimensional model representing the geometry of the eye comprises representations of a cornea, an anterior humor, a vitreous humor, a lens, a retina, a choroid, an iris, a ciliary body, and a sclera.

10. The system of claim 1, wherein the output device comprises a display.

11. A method for measuring a body temperature of a subject based on thermal data collected by an infrared camera, the method being performed using a processing unit comprising one or more processors and one or more memories storing at least one three-dimensional model representing a geometry of an eye and instructions configured to be executed by the processing unit, the three-dimensional model comprising a model ocular surface and a model back portion opposite the model ocular surface, the method comprising: receiving the thermal data collected by the infrared camera, the thermal data comprising a plurality of thermal measurements having respective pixel positions; determining an ocular surface portion of the thermal data that corresponds to a subject ocular surface of the subject; transitioning, based on the pixel positions, the thermal measurements of the ocular surface portion of the thermal data to locations on the model ocular surface on the three-dimensional model representing the geometry of the eye that correspond to locations on the subject ocular surface from which the thermal measurements were collected; and based on the thermal data, as transitioned to the model ocular surface of the three-dimensional model, determining a temperature at the model back portion of the three-dimensional model, wherein the determined temperature at the model back portion indicates the body temperature of the subject; and transmitting the determined temperature to an output device.

12. The method of claim 11, wherein determining the temperature at the model back portion of the three-dimensional model comprises calculating what the temperature at the model back portion of the three-dimensional model must be to produce the observed thermal data, as transitioned to the model ocular surface, based on known boundary conditions and environmental conditions.

13. The method of claim 12, wherein the environmental conditions include a convective heat transfer parameter of air adjacent to the subject.

14. The method of claim 11, wherein the method further comprises obtaining the thermal data using the infrared camera.

15. The method of claim 11, wherein determining an ocular surface portion of the thermal data is performed using an optical image of the subject that indicates a location of the subject ocular surface of the subject.

16. The method of claim 15, wherein the method further comprises processing the thermal data and/or the optical image to correct for differences in one or more of: (i) image sizes, (ii) orientations, or (iii) fields of view.

17. The method of claim 15, wherein the method further comprises obtaining the thermal data using the infrared camera and obtaining the optical image using an optical camera.

18. The method of claim 11, wherein the determined temperature indicates a core body temperature of the subject.

19. The method of claim 11, wherein the three-dimensional model representing the geometry of the eye comprises representations of a cornea, an anterior humor, a vitreous humor, a lens, a retina, a choroid, an iris, a ciliary body, and a sclera.

20. The method of claim 11, wherein the output device comprises a display.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The organization and manner of the structure and operation of the invention, together with further objects and advantages thereof, may best be understood by reference to the following description taken in connection with the accompanying drawings wherein like reference numerals identify like elements in which:

(2) FIG. 1 is a diagram of a system for estimating core body temperature using a thermal image according to one embodiment of the present disclosure.

(3) FIG. 2 is a thermal image prepared using the system of FIG. 1.

(4) FIG. 3 shows a flow chart of a method of estimating core body temperature according to one embodiment of the present disclosure.

(5) FIG. 4A is a diagram of the temperature data of the eye being applied to a model of an ocular surface of the eye according to one embodiment of the present disclosure.

(6) FIG. 4B is a diagram of the application of the temperature data being applied to the model of the eye surface according to one embodiment of the present disclosure.

(7) FIG. 5 is a diagram of an adult human eye model depicting the different regions considered in the formulating the mathematical model;

(8) FIG. 6A is a diagram of the boundary layers of the eye model as seen from the front according to one embodiment of the present disclosure;

(9) FIG. 6B is a diagram of the boundary layers of the eye model of FIG. 6A as a side cross-section.

DETAILED DESCRIPTION OF THE DISCLOSURE

(10) While this invention may be susceptible to embodiment in different forms, specific embodiments are shown in the drawings and will be described herein in detail with the understanding that the present disclosure is to be considered an exemplification of the principles of the invention and is not intended to limit the invention to that as illustrated.

(11) FIG. 1 shows a system 100 for monitoring the core body temperature of a human body remotely. The system includes optical (visible light) and infrared cameras 110 that can be used to acquire optical and thermal images of an ocular surface 160 of a human, respectively. The thermal image of the ocular surface may be a part of a thermal image of a facial surface 150 or of a thermal image of one or both eyes. In some embodiments, the optical image of one or both eyes may be derived from the image of the facial surface 150. The system 100 includes a processor 120 in communication with the optical and infrared cameras 110 and configured to receive the thermal image data. The processor 120 may be in communication with a memory 130 storing a program that, when executed, may perform a method to process the thermal image data. The memory 130 may be a non-transitory, computer readable medium, as would be understood by a person of ordinary skill in the art. The system 100 may also include an output device 140, such as a monitor, a transmitter, a memory, or a printer, for receiving information containing the core body temperature estimated by the processor 120 through the execution of the program on the memory 130. FIG. 2 shows a facial thermal image 250, corresponding to the facial surface 150, and ocular surface thermal images 260, corresponding to the ocular surfaces 160.

(12) FIG. 3 shows a method 300 of estimating a core body temperature according to one embodiment of the present disclosure. In step 310, at least part of the facial surface 150 including at least one ocular surface 160 may be detected by the optical and infrared cameras 110. In step 320, the facial thermal image 250 may be acquired by the infrared camera 110.

(13) In step 330, the facial thermal image 250 may be transmitted from the infrared camera 110 to the processor 120. In step 340, the processor 120, executing the program on the memory of a local machine or cloud computer 130, may extract the thermal image of the right eye 260r and/or the left eye 2601. Step 340 may be performed automatically or may be manually performed by a person controlling the isolation of the ocular surface thermal images 260 of one or both eyes. In some embodiments, step 340 may include facial feature detection where facial landmarks are identified to localize and isolate the eye regions 260 from the facial thermal image 250. The processor 120 may first identify the location of a face from the image 250 transferred from the infrared camera 110 using one or more of: Haar cascades, pre-trained Histogram of Oriented Gradients and Linear and support vector machine object detector, and deep learning-based algorithms. Once the location of the face is confirmed, then the processor 120 may execute a program for facial landmark detection to first identify eye regions, and, then, isolate the ocular surface thermal image 260 from the facial thermal image 250. The facial landmark detection may be based on pre-trained facial landmark images where ensemble regression trees are trained to estimate the facial landmark positions directly from pixel intensities. The facial landmark detection may be performed using a location of 68 (x, y) coordinate points iBUG 300-W dataset. The use of the iBug-W dataset is exemplary and illustrative only, as other suitable datasets known to persons of skill in the art may also be used, such as a 194 coordinate points HELEN dataset. The facial landmark detector may also be equipped with eye blink detection feature, where it can distinguish open eyes from closed once. The purpose of the blink feature is to track the eyes and facilitate automatic capture of the ocular surface images from open eyes from a recorded or live streaming infrared thermal image.

(14) In step 345, one of the optical image and the thermal image is mapped onto the other in order to correct for differences (image sizes, mounting orientations, fields of view, etc.) due to multiple cameras 110 being used to capture the optical and thermal images. This step is optional and may not be necessary when using a single camera that can capture both optical and infrared images. Once step 345 is complete, future performance of the method 300 may use the mapping information from previous performances, thus, step 345, if needed, is only performed once per set up of the cameras 110. Ocular landmark points (pixel locations) in the optical images may be used as a reference points to locate the ocular landmark points in the thermal images. This mapping process may use homography or geometric-based mapping.

(15) Homography mapping is performed by selecting 4-8 points in the optical image and the corresponding thermal image to set up an initial pixel correspondence between the two cameras 110. Then a homograph may be constructed based on those data points that can then be used to map pixel locations in the optical image onto the thermal image. Geometric-based mapping is performed by using the inner canthi locations in both the optical and thermal images as anchor points (or fixed points of reference). In the thermal image, the inner canthi are most consistently warmest part of the facial surface; hence, different methods can be used to locate the inner canthi, such as 1) a localized search of the pixels around the eye region with the warmest temperature values and 2) use of machine learning to identify location of the inner canthi. In the optical image, the location of the inner canthi may be obtained from the ocular landmark points. Following the localization of the inner canthi, the geometry of both images and the relationship between the image dimensions and Field of Views of the chosen cameras may then be used to map the ocular landmark points from the optical image onto the thermal image. Both homography and geometric based mapping require that the cameras 110 be in close proximity to each other during the taking of the optical and thermal images, usually between 0.5 and 2.5 inches, in order to reduce the effect of parallax.

(16) In step 350, temperature data may be extracted from the ocular surface thermal images 260 of the one or both eyes. Temperature data extraction may include filtering temperature values within segmented regions of the ocular surface 160.

(17) In step 360, a model 400 of the eye may be constructed with the considerations of different regions: cornea, anterior and vitreous humor, lens, retina, choroid, iris, ciliary body, and sclera, grouped based on their thermophysical properties. In some embodiments, step 360 may be optional, especially if the eye model 400 has been previously constructed or provided. The eye model 400 selected to be used to establish boundary conditions may differ based on the patient. Different sizes of eye models, ranging from an infant to an adult size, may be constructed using CAD and meshing software. The goal is to approximately match the eyeball sizes of different age groups for an accurate estimate of core body temperature. The globes of the eye models are assumed to be axisymmetric with respect to the optical axis. The anterior and posterior regions of the eye 500 are composed of vascular and avascular tissue structures (See FIG. 5), where the avascular part of the eye is composed of cornea 520, anterior chamber (aqueous humor) 530, lens 540, and vitreous humor 550, and the vascular structure 510 is composed of retina & choroid 514, iris & ciliary body 516, and sclera 512. Due to similarities in thermophysical properties, an eye model may be constructed by summarizing or using representative structures of the eye for arrive at an eye model using five regionsR.sub.1-R.sub.5 in Table 1. The thermophysical properties and other parameters of the eye model 400 are obtained from a wide range of literature and are expressed as statistical values (i.e., mean?SD).

(18) TABLE-US-00001 TABLE 1 Thermophysical properties of eye tissues (see FIG. 5) Thermal Eye conductivity, Specific heat, Density, ? Regions Tissue k [W/m-K] C.sub.p [J/kg-K] [kg/m.sup.3] R.sub.1 Sclera/Retina/ 1.0042 ? 0.04 4200 1032 Iris/Choroid R.sub.2 Cornea 0.58 ? 0.06 3515 1052 R.sub.3 Aqueous 0.58 ? 0.01 3997 1003 R.sub.4 Lens 0.43 ? 0.05 3133 1076 R.sub.5 Vitreous 0.59 ? 0.01 4047 1005 Blood 0.53 ? 0.04 3617 1050

(19) Temperature distribution within the eye may be formulated using an energy equation where the vascular region of the eye includes two energy equations, while the avascular region is treated as a solid tissue. In the vascular region 510 of the eye (R.sub.1), for blood:

(20) $\begin{array}{cc}?\left(?k_b?T_b\right)+h_{\mathrm{tb}}{a}_{\mathrm{tb}}\left(T_s-T_b\right)={\mathrm{??}}_b{c}_{p_b}\left(\frac{?T_b}{?t}+{\stackrel{.fwdarw.}{v}}_b?T_b\right),\mathrm{in}{R}_1& \left(1\right)\end{array}$
For tissue:

(21) $\begin{array}{cc}?.Math.\left(\left(1-?\right)k_s?T_s\right)+\left(1-?\right)Q_m+h_{\mathrm{tb}}{a}_{\mathrm{tb}}\left(T_b-T_s\right)=\left(1-?\right){?}_s{c}_{p_s}\frac{?T_s}{?t},\mathrm{in}{R}_1& \left(2\right)\end{array}$
In the avascular region 520, 530, 540, 550 of the eye (R.sub.i):

(22) $\begin{array}{cc}?.Math.\left(k_i?T_i\right)={?}_i{c}_{p_i}\frac{?T_i}{?t},\mathrm{in}{R}_i,i=2.Math.5& \left(3\right)\end{array}$

(23) where R.sub.1 and R.sub.i (i=2 . . . 5) represent the vascular and avascular regions (Table 1) in the eyeball, respectively; k.sub.i, k.sub.s and k.sub.b are the thermal conductivities of the avascular tissue (in R.sub.i, i=2 . . . 5), vascular tissue (in R.sub.I) and blood, respectively; ? is the porosity of R.sub.1; T.sub.b and T.sub.s are the temperature of the blood and tissue in R.sub.1, respectively; T.sub.i is the temperature of the avascular tissue in R.sub.i (i=2 . . . 5); h.sub.tb is the convective heat transfer coefficient between blood and tissue structure; a.sub.tb is the specific surface area; Q.sub.m is the volumetric heat generation rates due to metabolism; ?.sub.s, c.sub.p.sub.s and ?.sub.b, c.sub.p.sub.b are the densities and specific heats of the tissue and blood in R.sub.1, respectively; ?.sub.i and c.sub.p.sub.i are the densities and specific heats of the avascular tissue in R.sub.i (i=2 . . . 5), respectively; and {right arrow over (v)}.sub.b is the blood flow velocity.

(24) The governing equations are simplified by assuming steady state. For the vascular region, the two energy equations are coupled by the convective heat transfer term, h.sub.tb a.sub.tb (T.sub.b?T.sub.s):
?.Math.(?k.sub.b?T.sub.b)+?.Math.((1??)k.sub.s?T.sub.s)???.sub.bc.sub.p.sub.b{right arrow over (v)}.sub.b.Math.?T.sub.b+(1??)Q.sub.m=0, in R.sub.1(4)

(25) The above equation may be further simplified by assuming the capillary vessels within the vascular tissue are thermally insignificant; meaning, there is no temperature difference between tissue and blood, i.e. T.sub.s=T.sub.b, leading to:
{right arrow over (v)}.sub.b.Math.?T.sub.b??.sub.b(T.sub.s?T.sub.b)=0 in R.sub.1(5)

(26) where ?.sub.b is the perfusion rate of blood, defined as volume flow rate of blood per unit volume of the vascular tissue, in mL-blood/[(mL-tissue)-sec]. The temperatures T.sub.s and T.sub.b collectively are represented as T.sub.1 (T.sub.s=T.sub.b=T.sub.1). The vascular equation can then be represented as:
?.Math.((?k.sub.b?T.sub.1)+?.Math.((1??)k.sub.s?T.sub.1)?(1??)Q.sub.m=0, in R.sub.1(6)
?.Math.(?T.sub.1(?k.sub.b+(1??)k.sub.s))?(1??)Q.sub.m=0, in R.sub.1(7)
k.sub.eff=?k.sub.b+(1??)k.sub.s and Q.sub.eff=(1??)Q.sub.m(8)
?.Math.(k.sub.eff?T.sub.1)?Q.sub.eff=0, in R.sub.1(9)

(27) where k.sub.eff is the effective value of the thermal conductivity of eye tissue and Q.sub.eff is the effective volumetric heat generation rate due to metabolism

(28) The avascular region 520, 530, 540, 550 may be expressed as:
?.Math.(k.sub.1?T.sub.i)=0, in R.sub.i i=2 . . . 5(10)
Table 2 shows the design parameters considered in the eye model. Although some of the parameters, such as convective heat transfer coefficient of the blood, porosity, and emissivity, are assumed to be constant, the ambient temperature, convective heat transfer coefficient of the air are expected to vary with the environmental conditions. The parameter values shown in Table 2 are designed to represent eye model in a controlled indoor environment.

(29) TABLE-US-00002 TABLE 2 Values of parameters used in eye model Parameter Description/Unit Value h.sub.? Convective heat transfer coefficient 10 ? 2.5 between the tear film and the surrounding air [W/m.sup.2-K]* h.sub.b Convective heat transfer coefficient 110 ? 10 between the vascular region (R.sub.1) and the surrounding blood [W/m.sup.2-K] ? Stefan Boltzmann constant [W/m.sup.2-K4] 5.67 ? 10.sup.?8 ? Emissivity 0.975 ? 0.021 E.sub.v Tear evaporative heat flux [W/m.sup.2]* 40 ? 2.0 T.sub.? Ambient temperature [? C.]* 23 ? 1.0 ? Porosity of the healthy vascular tissue 0.3 ? 0.03 Q.sub.m Metabolic heat generation [W/m.sup.3] *Subject to change based on the environmental condition

(30) Combination the vascular and avascular equations will provide the mathematical formulation necessary for solving the core body temperature as the boundary condition of the sclera surface that is not exposed to the surrounding air.
?.Math.(k.sub.eff?T.sub.1).Math.?.Math.(k.sub.i?T.sub.i)?Q.sub.eff=0, in R.sub.i i=2 . . . 5(11)

(31) The boundary condition for the vascular surface of the sclera not exposed to the surrounding air 610 (See FIGS. 6A-6B), ?.sub.1 is:

(32) $\begin{array}{cc}-k_{\mathrm{eff}}\frac{?T_1}{?n_1}=h_b\left(T_1-T_{d\_\mathrm{core}}\right),\mathrm{on}{?}_1& \left(12\right)\end{array}$

(33) where h.sub.b is the convective heat transfer coefficient between the vascular region 510 (R.sub.1) and the surrounding blood, and

(34) $\frac{?T_1}{?n_1}$
is the vascular temperature gradient in the direction of outward unit vector (n.sub.1) normal to ?.sub.1 of R.sub.1. T.sub.d_core is the default core body temperature and T.sub.1 is the vascular tissue temperature distribution of R.sub.1.

(35) The boundary condition for the vascular surface of the sclera exposed to the surrounding air 620, ?.sub.2 is:

(36) $\begin{array}{cc}-k_{\mathrm{eff}}\frac{?T_1}{?n_1}=h_{?}\left(T_1-T_{?}\right)+??\left(T_1^4-T_{?}^4\right)+E_v,\mathrm{on}{?}_2& \left(13\right)\end{array}$

(37) The vascular surface 620 is subjected to both radiative and convective heat transfer, as well as evaporation from the tear film to the surrounding air.

(38) $\frac{?T_1}{?n_1}$
is the vascular temperature gradient in the direction of outward unit vector (n.sub.1) normal to ?.sub.2 of R.sub.1, h.sub.? is the convective heat transfer coefficient between the tear film and the surrounding air, ? is the Stefan-Boltzmann constant, ? is the ocular surface emissivity, T.sub.? is the ambient air temperature, E.sub.v is the tear evaporative heat flux. Values of the parameters used in the above equations, including the thermophysical properties, are summarized in Tables 1 and 2.

(39) The interfacial conditions between two contiguous regions are described by the continuity condition:

(40) $\begin{array}{cc}\begin{array}{c}{k}_{\mathrm{eff}}\frac{?T_1}{?n_1}=-k_i\frac{?T_i}{?n_i}\\ T_1=T_i\end{array}\},\mathrm{on}{\mathrm{IR}}_{1i}i=2.Math.5& \left(14\right)\end{array}$

(41) where IR.sub.1i is the interface between the vascular region R.sub.1 and the neighboring avascular regions R.sub.i (i=2 . . . 5); k.sub.i and T.sub.i are the thermal conductivity and the temperature distribution of R.sub.i (i=2 . . . 5), respectively.

(42) The boundary condition avascular surface 630 of the cornea ?.sub.3 (FIGS. 6A-6B) is:

(43) 0$\begin{array}{cc}-k_3\frac{?T_3}{?n_3}=h_{?}\left(T_3-T_{?}\right)+??\left(T_3^4-T_{?}^4\right)+E_v,\mathrm{on}{?}_3& \left(15\right)\end{array}$

(44) where k.sub.3 and T.sub.3 are the thermal conductivity and temperature of the cornea tissue of the avascular region R.sub.2, respectively,

(45) $\frac{?T_3}{?n_3}$
is the avascular temperature gradient in the direction of outward unit vector (n.sub.3) normal to ?.sub.3 of R.sub.2.

(46) $$\begin{gathered} \begin{array}{cc}{?}_R\left(k_{\mathrm{eff}}+k_i\right)?w.Math.?\mathrm{TdR}+{?}_{{?}_u={?}_2.Math.{?}_3}w\left[h_{?}\left(T-T_{?}\right)+?\left(T^4-T_{?}^4\right)+E_v\right]d{?}_u+ \\ {?}_{{?}_1}{\mathrm{wh}}_b\left(T-T_{d\_\mathrm{core}}\right)d{?}_1-{?}_R{\mathrm{wQ}}_{\mathrm{eff}}dR=0?w?H^{1}i=2.Math.5& \left(19\right)\end{array} \end{gathered}$$

(47) where the eye domain R?.sup.d the unknown temperature of the tissue T?C.sup.n (R) (where C.sup.n is a continuous function of n derivatives), and a test function w such that w?C.sup.n(R). Equation 19 is the final mathematical model that estimates the core body temperature.

(48) In some embodiments, an approximate solution to this mathematical model is computed numerically using techniques that may include fixed-point iteration, Newton's method, or a quasi-Newton method for handling the nonlinear radiative term, and techniques that may include direct solvers or preconditioned iterative solvers for systems of linear equations.

(49) In step 370, the temperature data may be transitioned from their pixel position in the ocular surface thermal image 260 to corresponding Cartesian coordinates 410 (see FIG. 4A) of a finite element mesh 400 of the eye model or another discretized model of the eye's geometry. The use of a Cartesian coordinate set is exemplary and illustrative only, as other coordinate systems may be used as would be understood by a person of ordinary skill in the art. In some embodiments, the transition may be performed using the Equation 20 for mapping:

(50) $$\begin{gathered} \begin{array}{cc}{Y}_{\mathrm{cor}}=1+\frac{\left(\left(P_y-Y_{\min }\right)*\left(N_{\mathrm{col}}-1\right)\right)}{Y_{\max }-Y_{\min }} \\ {Z}_{\mathrm{cor}}=1+\frac{\left(\left(P_Z-Z_{\min }\right)*\left(N_{\mathrm{row}}-1\right)\right)}{Z_{\max }-Z_{\min }}& \left(20\right)\end{array} \end{gathered}$$

(51) where N.sub.col and N.sub.row are the column and row numbers, respectively; P.sub.y and P.sub.z are the pixel locations in the column and row locations, respectively; Y.sub.min is the least-valued point of the y axis from the sensor data of the ocular surface; Y.sub.max is the most-valued point; Z.sub.min is the least-valued point of the z axis from the sensor data point of the ocular surface; Z.sub.max is the most-value point of the z axis from the sensor data point of the ocular surface; and Y.sub.cor and Z.sub.cor are the sensor locations in the eye model corresponding to the P.sub.y and P.sub.z pixels of the ocular surface thermal image 260.

(52) Zero or missing values may be validated. Once the temperature values are mapped from the 2D of the thermal image to 3D sensor data points of the ocular surface of the eye model 400, the eye model 400 may be electronically stored as an observed input value that may be used to estimate the temperature at the back of the eye using inverse analysis and optimization. In some embodiments, models of heat transfer may be used to map between core temperature and temperatures on the ocular surface.

(53) In step 380, the temperature at the back of the eye is estimated using the eye model 400. The estimate may be obtained using a gradient-based or non-gradient-based optimization procedure to estimate the core body temperature at the back of the eye that produces the ocular surface temperature profile that best fits the data according to some merit function. For the temperature estimation in step 380, the temperature data may be evaluated using a combination of forward and backward methods, otherwise known as inverse analysis. In one embodiment, the mathematical formulations of the eye model which may include a governing equation, formulated in Equation. 19, along with the goodness of fit test and non-gradient optimization are used to evaluate the surface temperature at the back of the eye. The inverse problem may be set up as a minimization function of temperature (T), where the analysis can be performed by iterative minimization of an object function. As an initial starting point, guess the value of T at the surface boundary of ?.sub.1 and solve the ocular surface temperature T at the surface boundary of ?.sub.2 and ?.sub.3 using a forward method. The estimated temperature data of the ocular surface (i.e. T at ?.sub.2 and ?.sub.3) may then undergo the goodness of fit test with the ocular surface temperature data obtained from the infrared camera where the convergence to the minimum of the objective function is assessed. Depending on the result of the fit, the initial guess may then be updated, and a new estimate of ocular surface temperature is evaluated. The iterative process continue until the objective function is satisfied, i.e. best core body temperature value is attained. To handle the iterative process, non-gradient-based optimization may be performed using Brent's method, golden section search method, downhill simplex, or a pattern search method; alternatively, a gradient-based technique such as steepest descent, conjugate gradient, the Broyden-Fletcher-Powell-Reeves method, or Newton's method may be used. The chi-square test may be used to assess the goodness of fit test between the estimated and infrared measured ocular surface temperature data.

(54) Equation 21 shows formulation of the reduced chi-square (?.sub.red.sup.2) equation, in which ?.sub.red.sup.2 is given by:

(55) $\begin{array}{cc}{?}_{\mathrm{red}}^2=\frac{1}{N}\underset{i=1}{\overset{N}{.Math.}}{\left(\frac{T_{\mathrm{IR}}\left(r_i;?\right)-T_i}{?\left(T\right)}\right)}^2& \left(21\right)\end{array}$

(56) where T.sub.IR represents the ocular surface temperature acquired from the thermal image; T.sub.i (where i=1 and 3) represents the estimated ocular surface temperature evaluated using Equation 19; r is the position of the temperature sensors, ? is the optimization parameter described as a function of the core body temperature; and N is the total number of sensors located on the ocular surface area of the eye model. Herein T.sub.IR(r; ?) is fitted to N number of numerically generated ocular surface temperature data (T.sub.i). Since ?.sub.red.sup.2 is formulated to account the noise inherited from the instrumentation error, (T) is added to account for the noise equivalence difference temperature of the camera provided by the manufacturer. It should be noted that, since (T) is uniform across the image, the method produces a core body temperature estimate that is independent of the accuracy rating of the camera.

(57) For a single model assessment, the accuracy of the fitness is established based on the following criteria: ?.sub.red.sup.2>1Poor fit. ?.sub.red.sup.2?1Good fit. ?.sub.red.sup.2<1Over fit.

(58) In step 390, the estimated core body temperature and (optionally) the model 400 may be transmitted to the output device 140.

(59) FIGS. 4A-4B show diagrams of temperature data being applied a model of the eye. In FIG. 4A, the thermal image of the right eye 260r is converted into coordinates 410 which relate to the finite element mesh 400. In FIG. 4B, the thermal image of the right eye 260r is shown being mapped to the mesh 400 which is used to form the eye model 420 with a back of the eye 430.

(60) While embodiments in the present disclosure have been described in some detail, according to the preferred embodiments illustrated above, it is not meant to be limiting to modifications such as would be obvious to those skilled in the art.

(61) The foregoing disclosure and description of the disclosure are illustrative and explanatory thereof, and various changes in the details of the illustrated apparatus and system, and the construction and the method of operation may be made without departing from the spirit of the disclosure.