Eye finding and tracking system

• US 5,859,686 A
• Filed: 05/19/1997
• Issued: 01/12/1999
• Est. Priority Date: 05/19/1997
• Status: Expired due to Term

• Current Assignee: MedFlex LLC
• Original Assignee: Northrop Grumman Corporation
• Inventors: Aboutalib, Omar, Ramroth, Richard Roy
• Primary Examiner(s): DANG, HUNG XUAN
• Application Number:
  US08/858,841
• Time in Patent Office: 603 Days
• Field of Search: 351/221, 351/211, 351/205, 351/212, 351/206, 351/246, 351/247, 351/210, 351/209
• US Class Current: 351/209
• CPC Class Codes: A61B 3/113 for determining or recording eye movement

Abstract

A system and method for finding and tracking the location of a subject'"'"'s eyes. This system and method employs an imaging apparatus which produces digital image frames including the face and eyes of a subject. An eye position finding and tracking apparatus is also included to average the intensity representing pixel values within respective M_x by M_y pixel blocks of a digitized image frame to create elements of output matrices. Then, elements of the output matrices are compared to various threshold values. These threshold values are chosen so as to identify which matrix elements correspond to a M_x by M_y pixel block potentially representing an image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris (i.e. a potential eye location). Once the flagged matrix elements have been determined, any M_x by M_y pixel block corresponding to a flagged matrix element can be designated as an actual subject eye location. However, it is preferred that further steps be taken to confirm the actual eye location, as well as to track the changing position of the subject'"'"'s eye. Accordingly, the preferred system and method include a provision for tracking the location of a center of each M_x by M_y pixel block representing a potential eye location in each subsequent image frame produced from the imaging apparatus. Further, a provision is included to detect a blink at the location. Such a blink confirms the potential eye location is an actual eye location.

Claims

1. A method of finding and tracking the location of a subject'"'"'s eyes, comprising the steps of:

employing an imaging apparatus which produces digital image frames including the face and eyes of a subject, each digital image frame comprising an array of pixel values representing the intensity of light reflected from the face of the subject, wherein each intensity representing pixel value is located at a position in the array specified by x and y coordinates;

averaging the intensity representing pixel values within respective M_x by M_y pixel blocks of a digitized image frame to create elements of output matrices; and

comparing elements of the output matrices to threshold values, said threshold values being chosen so as to identify which matrix elements correspond to a M_x by M_y pixel block which includes pixel values potentially representing an image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris.

2. The method of claim 1, wherein said averaging step comprises the steps of:

(a) respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks in the upper left-hand corner of said digitized image frame to create a first three-element column of an output matrix;

(b) respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks each of which is offset to the right by one column of the array in relation to a last averaged pixel block to create a new column of the output matrix; and

(c) repeating step (b) until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column of the array have been averaged to create a last column of the output matrix.

3. The method of claim 2, wherein said averaging step further comprises the steps of:

(d) respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks starting at a position on the left-hand side of the array which is offset downward by one row in relation to a previously averaged pixel block to create a first three-element column of a new output matrix;

(e) respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks each of which is offset to the right by one column of the array in relation to a last averaged pixel block to create a new column of the new output matrix;

(f) repeating step (e) until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column of the array have been averaged to create a last column of the new output matrix; and

(g) repeating steps (d) through (f) until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column and row of the array have been averaged to create a last column of a last new output matrix.

4. The method of claim 3, wherein said comparing step comprises the steps of:

(h) comparing each element of each output matrix to a threshold range;

(i) flagging any output matrix element which both exceeds a lower limit of the threshold range and is less than an upper limit of the threshold range;

(j) comparing a threshold value to the average of the intensity representing pixel values within each M_x by M_y pixel block immediately surrounding any pixel block corresponding to an output matrix element flagged in step (i).

5. The method of claim 4, wherein said threshold range comprises:

a lower limit representing the lowest expected average of the intensity representing pixel values within a M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced; and

an upper limit representing the highest expected average of the intensity representing pixel values within the M_x by M_y pixel block corresponding to the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced.

6. The method of claim 4, wherein the same threshold value is compared to the average of the intensity representing pixel values associated with each M_x by M_y pixel block immediately surrounding any pixel block corresponding to a flagged output matrix element, and wherein said threshold value comprises a lowest expected average of the intensity representing pixel values within any M_x by M_y pixel block immediately surrounding a M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced, and wherein the comparing step further comprises:

(k) de-flagging any previously flagged output matrix element which has any surrounding pixel block which fails to exceed the threshold value.

7. The method of claim 4, wherein a different threshold value is compared to the average of the intensity representing pixel values associated with each M_x by M_y pixel block immediately surrounding any pixel block corresponding to a flagged output matrix element, and wherein the respective threshold value comprises a lowest expected average of the intensity representing pixel values within the associated M_x by M_y pixel block immediately surrounding a central M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced in immediately surrounding pixel blocks which exhibit an average intensity exceeding that of an average intensity of said digitized image frame, and a highest expected average of the intensity representing pixel values within the associated M_x by M_y pixel block immediately surrounding the central M_x by M_y pixel block corresponding to the image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced in immediately surrounding pixel blocks which exhibit an average intensity less than that of the average intensity of said digitized image frame, and wherein the comparing step further comprises:

(k) de-flagging any previously flagged output matrix element which has any surrounding pixel block which either (1) fails to exceed the threshold value in the immediately surrounding pixel blocks which exhibit an average intensity exceeding that of an average intensity of said digitized image frame, or (2) fails to be less than the threshold value produced in the immediately surrounding pixel blocks which exhibit an average intensity less than that of the average intensity of said digitized image frame.

8. The method of claim 4, further comprising the steps of:

(l) identifying groups of flagged matrix elements having any common intensity representing pixel values associated with the respective M_x by M_y pixel blocks corresponding to the flagged matrix elements;

(m) determining one of (1) which of the flagged matrix elements in each of said groups identified in step (l) are greater than the others, or (2) whenever no one element has the greatest value, which of the flagged matrix elements in each of said groups identified in step (l) is in the geometric center of the group; and

(n) deflagging any of the flagged matrix elements identified in step (l) which were not determined to be greater or in the center in step (m).

9. The method of claim 8, further comprising the step of determining if a M_x by M_y pixel block corresponding to a flagged matrix element is within prescribed horizontal and vertical distance ranges from a M_x by M_y pixel block corresponding to another flagged matrix element, said prescribed horizontal distance range having an upper limit corresponding to the maximum expected horizontal eye separation of the subject and a lower limit corresponding to the minimum expected horizontal eye separation of the subject, and said prescribed vertical distance range having an upper limit corresponding to the maximum expected vertical eye separation of the subject and a lower limit corresponding to the minimum expected vertical eye separation of the subject.

10. The method of claim 9, further comprising the step of designating any M_x by M_y pixel block corresponding to a flagged matrix element as an actual eye location.

11. The method of claim 8, further comprising the step of designating any M_x by M_y pixel block corresponding to a flagged matrix element as an actual eye location.

12. A method of finding and tracking the location of a subject'"'"'s eyes, comprising the steps of:

employing an imaging apparatus which produces digital image frames including the face and eyes of a subject, each digital image frame comprising an array of pixel values representing the intensity of light reflected from the face of the subject, wherein each intensity representing pixel value is located at a position in the array specified by x and y coordinates;

averaging the intensity representing pixel values within respective M_x by M_y pixel blocks of a first digitized image frame to create elements of output matrices;

comparing elements of the output matrices associated with the first image frame to threshold values, said threshold values being chosen so as to identify which matrix elements correspond to a M_x by M_y pixel block which includes pixel values potentially representing an image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris; and

tracking the location of a center of each M_x by M_y pixel block potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in each subsequent image frame produced from the imaging apparatus, and detecting a blink at the location, said blink indicating the location is an actual eye location.

13. The method of claim 12, wherein the tracking and blink detecting step comprises the step of:

(a) determining the center of the M_x by M_y pixel block which includes pixel values potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris.

14. The method of claim 13, wherein the tracking and blink detecting step further comprises the steps of:

(b) selecting cut-out pixel blocks in a next consecutive image frame produced by the imaging apparatus, each of said cut-out pixel blocks respectively comprising a M_x by M_y pixel block corresponding to the location of a pixel block in a immediately preceding image frame which was identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris, as well as a number of pixels which surround the identified pixel block;

(c) respectively forming a matrix of intensity representing pixel values from an area surrounding the center of each M_x by M_y pixel block identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in a last preceding image frame in which such a center was determined;

(d) correlating an appropriate one of the matrices formed in step (c) with each element of the associated cutout block, said correlation being performed by first sequentially overlaying the center element of said appropriate one of said matrices onto each pixel value of said associated cut-out block starting with the uppermost left-hand corner, and second, performing a correlation between the overlaid matrix and the cut-out block for each overlaid cut-out block pixel location, to create a matrix of correlation coefficients for each cut-out block in said next consecutive image frame;

(e) comparing a correlation threshold value to each element in each correlation coefficient matrix;

(f) flagging any correlation coefficient matrix element which exceeds said correlation threshold value in each of the correlation coefficient matrices;

(g) if elements of a correlation coefficient matrix are flagged in step (f), determining which flagged correlation coefficient element in the correlation coefficient matrix has the greatest value, wherein said element having the greatest value in the matrix corresponds to the center pixel value of the M_x by M_y pixel block which includes pixel values potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in said next consecutive image frame;

(h) if no elements of a correlation coefficient matrix are flagged in step (f), noting the lack of flagged elements condition in the matrix and calculating the number of consecutive times this condition has occurred; and

(i) repeat steps (b) through (h) for each image frame produced by the imaging apparatus.

15. The method of claim 14, wherein each of said cut-out pixel blocks constitutes the M_x by M_y pixel block corresponding to the location of the pixel block in the immediately preceding image frame which was identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris, and all surrounding M_x by M_y pixel blocks which are adjacent to said pixel block.

16. The method of claim 14, wherein said area referred to in step (c) constitutes the M_x by M_y pixel block identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in the last preceding image frame in which said center was determined, and all surrounding M_x by M_y pixel blocks which are adjacent to the identified pixel block.

17. The method of claim 14, wherein said threshold value is chosen so as to ensure a substantial degree of correlation between an overlaid matrix and an associated cut-out block.

18. The method of claim 14, wherein the tracking and blink detecting step further comprises the steps of:

(j) detecting that a blink has occurred in a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever the number of consecutive times the lack of flagged elements condition calculated in step (h) for a correlation coefficient matrix associated with said location does not exceed about seven before elements of said correlation coefficient matrix are flagged in step (f);

(k) designating the center of the M_x by M_y pixel block corresponding said location where a blink has been detected in step (j) as an actual eye location.

19. The method of claim 18, further comprising the steps of:

assigning a low confidence status to a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever the number of consecutive times the lack of flagged elements condition calculated in step (h) for a correlation coefficient matrix associated with said location exceeds 0.5, said low confidence status indication the location is not likely to be an actual eye; and

assigning a low confidence status to a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever a blink is not detected in step (j) at the location for 150 consecutive image frames.

20. The method of claim 19, further comprising the step of assigning a low confidence status to a location identified as an actual eye location in step (k) whenever the number of consecutive times the lack of flagged elements condition calculated in step (h) for a correlation coefficient matrix associated with said location exceeds 150 and there is no other location identified as an actual eye location within prescribed horizontal and vertical distance ranges from the actual eye location, said prescribed horizontal distance range having an upper limit corresponding to the maximum expected horizontal eye separation of the subject and a lower limit corresponding to the minimum expected horizontal eye separation of the subject, and said prescribed vertical distance range having an upper limit corresponding to the maximum expected vertical eye separation of the subject and a lower limit corresponding to the minimum expected vertical eye separation of the subject.

21. The method of claim 20, further comprising the step of reinitializing the system whenever all the locations identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris or actual eye locations are assigned a low confidence status.

22. An eye position finding and tracking system comprising:

an imaging apparatus which produces digital image frames including the face and eyes of a subject, each digital image frame comprising an array of pixel values representing the intensity of light reflected from the face of the subject, wherein each intensity representing pixel value is located at a position in the array specified by x and y coordinates; and

a eye finding and tracking unit comprising;

a first processor portion capable of averaging the intensity representing pixel values within respective M_x by M_y pixel blocks of a digitized image frame to create elements of output matrices,

a second processor portion capable of comparing elements of the output matrices to threshold values, said threshold values being chosen so as to identify which matrix elements correspond to a M_x by M_y pixel block which includes pixel values potentially representing an image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris.

23. The system of claim 22, wherein said first processor portion comprises:

first means for respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks in the upper left-hand corner of said digitized image frame to create a first three-element column of an output matrix;

second means for respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks each of which is offset to the right by one column of the array in relation to a last averaged pixel block to create a new column of the output matrix; and

first repeating means for repeatedly employing the first and second averaging means until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column of the array have been averaged to create a last column of the output matrix.

24. The system of claim 23, wherein said first processor portion further comprises:

third means for respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks starting at a position on the left-hand side of the array which is offset downward by one row in relation to a previously averaged pixel block to create a first three-element column of a new output matrix;

fourth means for respectively averaging the intensity representing pixel values within three vertically arranged M_x by M_y pixel blocks each of which is offset to the right by one column of the array in relation to a last averaged pixel block to create a new column of the new output matrix;

second repeating means for repeatedly employing the third and fourth averaging means until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column of the array have been averaged to create a last column of the new output matrix; and

third repeating means for repeatedly employing the third and fourth averaging means and the second repeating means until three vertically arranged M_x by M_y pixel blocks which include pixel values from the last column and row of the array have been averaged to create a last column of a last new output matrix.

25. The system of claim 24, wherein said second processor portion comprises:

first means for comparing each element of each output matrix to a threshold range;

means for flagging any output matrix element which both exceeds a lower limit of the threshold range and is less than an upper limit of the threshold range;

second means for comparing a threshold value to the average of the intensity representing pixel values within each M_x by M_y pixel block immediately surrounding any pixel block corresponding to an output matrix element flagged by the flagging means.

26. The system of claim 25, wherein said threshold range comprises:

a lower limit representing the lowest expected average of the intensity representing pixel values within a M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced; and

an upper limit representing the highest expected average of the intensity representing pixel values within the M_x by M_y pixel block corresponding to the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced.

27. The system of claim 25, wherein the same threshold value is compared to the average of the intensity representing pixel values associated with each M_x by M_y pixel block immediately surrounding any pixel block corresponding to a flagged output matrix element, and wherein said threshold value comprises a lowest expected average of the intensity representing pixel values within any M_x by M_y pixel block immediately surrounding a M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced, and wherein the second processor portion further comprises means for de-flagging any previously flagged output matrix element which has any surrounding pixel block which fails to exceed the threshold value.

28. The system of claim 27, wherein said eye finding and tracking unit further comprises a third processor portion comprising:

means for identifying groups of flagged matrix elements having any common intensity representing pixel values associated with the respective M_x by M_y pixel blocks corresponding to the flagged matrix elements;

means for determining one of (1) which of the flagged matrix elements in each of said groups are greater than the others, or (2) whenever no one element has the greatest value, which of the flagged matrix elements in each of said groups is in the geometric center of the group; and

second means for deflagging any of the flagged matrix elements identified by the identifying means which were not determined to be greater or in the center by the determining means.

29. The system of claim 28, wherein said eye finding and tracking unit further comprises a fourth processor portion capable of determining if a M_x by M_y pixel block corresponding to a flagged matrix element is within prescribed horizontal and vertical distance ranges from a M_x by M_y pixel block corresponding to another flagged matrix element, said prescribed horizontal distance range having an upper limit corresponding to the maximum expected horizontal eye separation of the subject and a lower limit corresponding to the minimum expected horizontal eye separation of the subject, and said prescribed vertical distance range having an upper limit corresponding to the maximum expected vertical eye separation of the subject and a lower limit corresponding to the minimum expected vertical eye separation of the subject.

30. The system of claim 25, wherein a different threshold value is compared to the average of the intensity representing pixel values associated with each M_x by M_y pixel block immediately surrounding any pixel block corresponding to a flagged output matrix element, and wherein the respective threshold value comprises a lowest expected average of the intensity representing pixel values within the associated M_x by M_y pixel block immediately surrounding a central M_x by M_y pixel block corresponding to an image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced in immediately surrounding pixel blocks which exhibit an average intensity exceeding that of an average intensity of said digitized image frame, and a highest expected average of the intensity representing pixel values within the associated M_x by M_y pixel block immediately surrounding the central M_x by M_y pixel block corresponding to the image of the subject'"'"'s pupil and at least a portion of the subject'"'"'s iris for the particular illumination condition present at the time the image was produced in immediately surrounding pixel blocks which exhibit an average intensity less than that of the average intensity of said digitized image frame, and wherein the second processor portion further comprises means for de-flagging any previously flagged output matrix element which has any surrounding pixel block which either (1) fails to exceed the threshold value in the immediately surrounding pixel blocks which exhibit an average intensity exceeding that of an average intensity of said digitized image frame, or (2) fails to be less than the threshold value produced in the immediately surrounding pixel blocks which exhibit an average intensity less than that of the average intensity of said digitized image frame.

31. The system of claim 30, wherein said eye finding and tracking unit further comprises a third processor portion comprising:

means for identifying groups of flagged matrix elements having any common intensity representing pixel values associated with the respective M_x by M_y pixel blocks corresponding to the flagged matrix elements;

means for determining one of (1) which of the flagged matrix elements in each of said groups are greater than the others, or (2) whenever no one element has the greatest value, which of the flagged matrix elements in each of said groups is in the geometric center of the group; and

second means for deflagging any of the flagged matrix elements identified by the identifying means which were not determined to be greater or in the center by the determining means.

32. The system of claim 31, wherein said eye finding and tracking unit further comprises a fourth processor portion capable of determining if a M_x by M_y pixel block corresponding to a flagged matrix element is within prescribed horizontal and vertical distance ranges from a M_x by M_y pixel block corresponding to another flagged matrix element, said prescribed horizontal distance range having an upper limit corresponding to the maximum expected horizontal eye separation of the subject and a lower limit corresponding to the minimum expected horizontal eye separation of the subject, and said prescribed vertical distance range having an upper limit corresponding to the maximum expected vertical eye separation of the subject and a lower limit corresponding to the minimum expected vertical eye separation of the subject.

33. An eye position finding and tracking system comprising:

an imaging apparatus which produces digital image frames including the face and eyes of a subject, each digital image frame comprising an array of pixel values representing the intensity of light reflected from the face of the subject, wherein each intensity representing pixel value is located at a position in the array specified by x and y coordinates; and

a eye finding and tracking unit comprising;

a first processor portion capable of averaging the intensity representing pixel values within respective M_x by M_y pixel blocks of a first digitized image frame to create elements of output matrices,

a second processor portion capable of comparing elements of the output matrices associated with the first image frame to threshold values, said threshold values being chosen so as to identify which matrix elements correspond to a M_x by M_y pixel block which includes pixel values potentially representing an image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris,

a third processor portion capable of tracking the location of a center of each M_x by M_y pixel block potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in each subsequent image frame produced from the imaging apparatus, and detecting a blink at the location, said blink indicating the location is an actual eye location.

34. The system of claim 33, wherein the third processor portion comprises first determining means for determining the center of the M_x by M_y pixel block which includes pixel values potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris.

35. The system of claim 34, wherein the third processor portion further comprises:

means for selecting cut-out pixel blocks in a next consecutive image frame produced by the imaging apparatus, each of said cut-out pixel blocks respectively comprising a M_x by M_y pixel block corresponding to the location of a pixel block in a immediately preceding image frame which was identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris, as well as a number of pixels which surround the identified pixel block;

means for respectively forming a matrix of intensity representing pixel values from an area surrounding the center of each M_x by M_y pixel block identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in a last preceding image frame in which such a center was determined;

means for correlating an appropriate one of the matrices formed by the forming means with each element of the associated cutout block, said correlation being performed by first sequentially overlaying the center element of said appropriate one of said matrices onto each pixel value of said associated cut-out block starting with the uppermost left-hand corner, and second, performing a correlation between the overlaid matrix and the cut-out block for each overlaid cut-out block pixel location, to create a matrix of correlation coefficients for each cut-out block in said next consecutive image frame;

means for comparing a correlation threshold value to each element in each correlation coefficient matrix;

means for flagging any correlation coefficient matrix element which exceeds said correlation threshold value in each of the correlation coefficient matrices;

second determining means for determining which flagged correlation coefficient element in the correlation coefficient matrix has the greatest value whenever elements of a correlation coefficient matrix are flagged by the flagging means, wherein said element having the greatest value in the matrix corresponds to the center pixel value of the M_x by M_y pixel block which includes pixel values potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in said next consecutive image frame;

means for noting the lack of flagged elements condition in the matrix and calculating the number of consecutive times this condition has occurred whenever no elements of a correlation coefficient matrix are flagged by the flagging means; and

means for repeatedly employing the selecting, forming, correlating, comparing flagging second determining, and noting means for each image frame produced by the imaging apparatus.

36. The system of claim 35, wherein each of said cut-out pixel blocks constitutes the M_x by M_y pixel block corresponding to the location of the pixel block in the immediately preceding image frame which was identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris, and all surrounding M_x by M_y pixel blocks which are adjacent to said pixel block.

37. The system of claim 35, wherein said area surrounding the center of each M_x by M_y pixel block identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in the last preceding image frame in which such a center was determined constitutes the M_x by M_y pixel block identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris in the last preceding image frame in which said center was determined, and all surrounding M_x by M_y pixel blocks which are adjacent to the identified pixel block.

38. The system of claim 35, wherein said threshold value is chosen so as to ensure a substantial degree of correlation between an overlaid matrix and an associated cut-out block.

39. The system of claim 35, wherein the third processor portion further comprises:

means for detecting that a blink has occurred in a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever the number of consecutive times the lack of flagged elements condition calculated by the noting means for a correlation coefficient matrix associated with said location does not exceed about seven before elements of said correlation coefficient matrix are flagged by the flagging means;

means for designating the center of the M_x by M_y pixel block corresponding said location where a blink has been detected by the detecting means as an actual eye location.

40. The system of claim 39, further comprising a fourth processor portion, said fourth processor portion comprising:

first means for assigning a low confidence status to a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever the number of consecutive times the lack of flagged elements condition calculated by the noting means for a correlation coefficient matrix associated with said location exceeds 150, said low confidence status indication the location is not likely to be an actual eye; and

second means for assigning a low confidence status to a location identified as potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris whenever a blink is not detected by the detecting means at the location for 150 consecutive image frames.

41. The system of claim 40, wherein the fourth processor portion further comprises third means for assigning a low confidence status to a location identified as an actual eye location by the designating means whenever the number of consecutive times the lack of flagged elements condition calculated by the noting means for a correlation coefficient matrix associated with said location exceeds 150 and there is no other location identified as an actual eye location within prescribed horizontal and vertical distance ranges from the actual eye location, said prescribed horizontal distance range having an upper limit corresponding to the maximum expected horizontal eye separation of the subject and a lower limit corresponding to the minimum expected horizontal eye separation of the subject, and said prescribed vertical distance range having an upper limit corresponding to the maximum expected vertical eye separation of the subject and a lower limit corresponding to the minimum expected vertical eye separation of the subject.

42. The system of claim 41, wherein the fourth processor portion further comprises means for reinitializing the system whenever all the locations identified as (i) potentially representing the image of the subject'"'"'s pupil and at least the portion of the subject'"'"'s iris and (ii) actual eye locations are assigned a low confidence status.