System and method for perceived image processing in a gaze tracking system

1. A system for measuring and modifying at least one model parameter of an object of an image in order to distinguish the object from noise in the image, comprising:

a perceived image generator, wherein the perceived image generator produces a first perceived image of the object based on a priori knowledge of one or more physical and/or optical properties of the object and the at least one model parameter that represents a measurement of at least one of the one or more physical and/or optical properties of the object;

an image-match function, wherein the image-match function compares the first perceived image with a real image of the object; and

a parameter adjustment function, wherein the parameter adjustment function adjusts the at least one model parameter so that the perceived image generator produces a second perceived image of the object that more closely matches the real image than the first perceived image.

2. A method for measuring and modifying at least one model parameter of an object of an image in order to distinguish the object from noise in the image, comprising:

producing a first perceived image of the object based on a priori knowledge of one or more physical and/or optical properties of the object and the at least one model parameter that represents a measurement of at least one of the one or more physical and/or optical properties of the object;

comparing the first perceived image with a real image of the object; and

adjusting the at least one model parameter so that the perceived image generator produces a second perceived image of the object that more closely matches the real image than the first perceived image.

3. The system of claim 1 , wherein the image-match function performs a pixel-by-pixel subtraction means to subtract intensities of pixels of the perceived-image from intensities of pixels of the real image to form a residual image, wherein the residual image pixels intensities have both positive and negative values, and computes an image-match metric as an overall magnitude of the residual image.

4. The system of claim 3 , wherein the image-match metric is computed as a sum of absolute values of the residual image.

5. The system of claim 3 , wherein the image-match metric is computed as an energy or root-mean-square of the residual image intensity profile.

6. The system of claim 3 , wherein the parameter adjustment means finds partial derivatives of the residual image magnitude image-match metric with respect to the at least one model parameter, and estimates a set of parameter corrections based upon a regression of the partial derivatives.

7. The system of claim 1 , wherein the object comprises an eye.

8. The system of claim 7 , wherein the one or more physical and/or optical properties of the object comprise a pupil of the eye.

9. The system of claim 7 , wherein the one or more physical and/or optical properties of the object comprise a corneal reflection from the eye.

10. The system of claim 1 , wherein

the real image is captured by a camera with an asymmetric aperture as an out-of-focus image,

a well-focused image is generated using the at least one model parameter comprising a focus condition of the real image, and

the perceived image is constructed as a convolution of an aperture transform and the well-focused image.

11. The method of claim 2 , wherein comparing the first perceived image with a real image of the object comprises performing a pixel-by-pixel subtraction means to subtract intensities of pixels of the perceived-image from intensities of pixels of the real image to form a residual image, wherein the residual image pixels intensities have both positive and negative values, and computing an image-match metric as an overall magnitude of the residual image.

12. The method of claim 11 , wherein the image-match metric is computed as a sum of absolute values of the residual image.

13. The method of claim 11 , wherein the image-match metric is computed as an energy or root-mean-square of the residual image intensity profile.

14. The method of claim 11 , wherein adjusting the at least one model parameter comprises finding partial derivatives of the residual image magnitude image-match metric with respect to the at least one model parameter, and estimating a set of parameter corrections based upon a regression of the partial derivatives.

15. The method of claim 2 , wherein the object comprises an eye.

16. The method of claim 15 , wherein the one or more physical and/or optical properties of the object comprise a pupil of the eye.

17. The method of claim 15 , wherein the one or more physical and/or optical properties of the object comprise a corneal reflection from the eye.

18. The method of claim 2 , wherein

the real image is captured by a camera with an asymmetric aperture as an out-of-focus image,

a well-focused image is generated using the at least one model parameter comprising a focus condition of the real image, and

the perceived image is constructed as a convolution of an aperture transform and the well-focused image.

19. The method of claim 2 , wherein the at least one model parameter comprises one of corneal-reflection (glint) coordinates, corneal-reflection intensity, pupil coordinates, pupil diameter, pupil intensity, or iris intensity.

20. The method of claim 2 , wherein the at least one model parameter comprises a focus-offset condition of a camera lens.