Systems and method for automatically choosing visual characteristics to highlight a target against a background

• US 20040120574A1
• Filed: 12/18/2002
• Published: 06/24/2004
• Est. Priority Date: 12/18/2002
• Status: Active Grant

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Rosenholtz, Ruth E.
• Granted Patent: US 7,130,461 B2
• Time in Patent Office: Days
• US Class Current: 382/165
• CPC Class Codes: G06T 11/60 Editing figures and text; Combining figures or text
• CPC Class Codes: G06T 2207/30176 Document
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields
• CPC Class Codes: G06V 30/18105 related to colour

Abstract

A method for choosing visual characteristics of a characteristic space within a region includes calculating a mean characteristic of the region, determining a covariance of the region, inverting the covariance to produce an inverse covariance, estimating a target characteristic, determining a difference between the mean characteristic and the target characteristic transposing the difference to produce a difference transpose, determining a saliency from a product of the difference transpose multiplied by the inverse covariance multiplied by the difference, comparing the saliency to an acceptance criterion to yield one of an acceptance if the saliency satisfies the acceptance criterion and a rejection otherwise, adjusting the target characteristic and repeating the determining a difference, transposing the difference, determining a saliency and comparing the saliency steps if the comparing yields the rejection, and outputting the target characteristic if the comparing yields the acceptance.

Claims

1. A method for choosing visual characteristics of a characteristic space within a region, comprising:

determining a saliency based on an inverse covariance of the region and a difference between a target characteristic and a mean characteristic of the region;

comparing the saliency against at least one acceptance criterion;

adjusting the target characteristic and repeating the determining the saliency and comparing the saliency steps if the saliency does not satisfy the at least one acceptance criterion; and

outputting the target characteristic if the saliency satisfies the at least one acceptance criterion.

2. The method of claim 1, wherein the visual characteristics are colors and the characteristic space is a color space.

3. The method of claim 1, wherein determining a saliency comprises:

determining a mean characteristic of the region;

determining a covariance of the region;

inverting the covariance to produce an inverse covariance;

estimating a target characteristic;

determining a difference between the mean characteristic and the estimated target characteristic;

transposing the difference to produce a difference transpose; and

determining a product of the difference transpose multiplied by the inverse covariance multiplied by the difference to produce the saliency.

4. The method of claim 3, wherein determining a covariance comprises determining a region distance between separate portions of the region.

5. The method of claim 4, wherein determining a covariance comprises determining at least one of a maximum distance and a Euclidean distance between the separate portions of the region.

6. The method of claim 3, wherein determining the mean characteristic comprises determining at least one of an arithmetic mean, a median, a mode and a range midpoint.

7. The method of claim 1, wherein determining a saliency comprises determining a Mahalanobis distance.

8. The method of claim 7, wherein determining a Mahalanobis distance comprises determining a sequence of:

Δ

²=(T-μ

)′

Σ

<sup>−

1</sup>(T-μ

), wherein

Δ

is the Mahalanobis distance,

T is the target characteristic,

μ

is the mean characteristic, and

Σ

<sup>−

1</sup> is the inverse covariance.

9. The method of claim 1, wherein the characteristic space has a single coordinate, and the mean characteristic, the covariance and the target characteristic each have a scalar element.

10. The method of claim 1, wherein the characteristic space has a plurality of coordinates, the mean characteristic is a mean vector having mean vector elements, the covariance is a matrix having tensor elements, and the target characteristic is a target vector having target vector elements.

11. The method of claim 10, wherein determining the saliency includes multiplying a weighting factor to the mean characteristic in each vector element for each coordinate of the plurality of coordinates.

12. The method of claim 10, wherein determining the saliency includes adding a bias factor to the mean characteristic in each vector element for each coordinate of the plurality of coordinates.

13. The method of claim 10, wherein estimating the target characteristic comprises subtracting a vector element for a coordinate of the mean characteristic from a midpoint of the coordinate.

14. The method of claim 3, wherein determining the mean characteristic of the region includes determining a mean characteristic of an object within the region.

15. The method of claim 1, wherein adjusting the target characteristic comprises incrementally changing a vector element of the target characteristic.

16. The method of claim 15, wherein incrementally changing a vector element of the target characteristic includes increasing an element difference between the vector element of the target characteristic and a vector element of the mean characteristic for a coordinate of the plurality of coordinates.

17. The method of claim 10, wherein adjusting the target characteristic comprises incrementally changing a vector element of the target characteristic.

18. The method of claim 1, wherein outputting the target characteristic comprises adjusting a characteristic of a target to match the target characteristic.

19. The method of claim 1, wherein outputting the target characteristic comprises displaying the target characteristic as a recommendation for a target.

20. The method of claim 19, wherein displaying the target characteristic comprises displaying in a format of at least one of a text, a set of vector elements in the characteristic space, a characteristic map, or an area containing the target characteristic.

21. The method of claim 1, further comprising outputting one of either the target characteristic or a status message if the saliency does not satisfy the at least one acceptance criterion.

22. The method of claim 1, wherein

adjusting the at least one acceptance criterion comprises adjusting the at least one acceptance criterion if the saliency does not satisfy the at least one acceptance criterion;

and

outputting the target characteristic comprises outputting the target characteristic if the saliency satisfies all of the at least one acceptance criterion.

23. The method of claim 1, wherein

adjusting the target characteristic comprises adjusting the target characteristic if the saliency does not satisfy all of the at least one acceptance criterion;

and

outputting the target characteristic if the saliency satisfies all of the at least one acceptance criterion.

24. The method of claim 1, wherein outputting the target characteristic comprises outputting the target characteristic if a nontarget characteristic satisfies at least one nontarget acceptance criterion.

25. A method for choosing characteristics of a characteristic space within a region, comprising:

determining a mean characteristic of the region;

determining a covariance of the region;

inverting the covariance to produce an inverse covariance;

estimating a target characteristic;

determining a difference between the mean characteristic and the target characteristic;

transposing the difference to produce a difference transpose;

determining a saliency from a product of the difference transpose multiplied by the inverse covariance multiplied by the difference;

comparing the saliency against at least one acceptance criterion;

adjusting the target characteristic and repeating the determining the saliency and comparing the saliency steps if the saliency does not satisfy the at least one acceptance criterion; and

outputting the target characteristic if the saliency satisfies the at least one acceptance criterion.

26. A computer storage medium having executable software code, the executable software code including:

instructions for determining a saliency based on an inverse covariance of the region and a difference between a target characteristic and a mean characteristic of the region;

instructions for comparing the saliency against at least one acceptance criterion;

instructions for adjusting the target characteristic and repeating the determining the saliency and comparing the saliency steps if the saliency does not satisfy the at least one acceptance criterion; and

instructions for outputting the target characteristic if the saliency satisfies the at least one acceptance criterion.

27. The computer storage medium of claim 26, wherein the instructions for determining the saliency further include:

instructions for determining a mean characteristic of the region;

instructions for determining a covariance of the region;

instructions for inverting the covariance to produce an inverse covariance;

instructions for estimating a target characteristic;

instructions for determining a difference between the mean characteristic and the estimated target characteristic;

instructions for transposing the difference to produce a difference transpose; and

instructions for determining a product of the difference transpose multiplied by the inverse covariance multiplied by the difference.

28. The computer storage medium of claim 26, wherein the visual characteristics are colors and the characteristic space is a color space.

29. The computer storage medium of claim 26, wherein the characteristic space has a plurality of coordinates, the mean characteristic is a vector having vector elements, the covariance is a matrix having tensor elements, and the target characteristic is a vector having vector elements.

30. The computer storage medium of claim 27, wherein the instructions for estimating the target characteristic comprise instructions for subtracting a vector element for a coordinate of the mean characteristic from a midpoint of the coordinate.

31. The computer storage medium of claim 26, wherein the instructions for determining a saliency comprise instructions for determining a Mahalanobis distance.

32. The computer storage medium of claim 31, wherein the instructions for determining a Mahalanobis distance comprise instructions for determining a sequence of:

Δ

²=(T-μ

)′

Σ

<sup>−

1</sup>(T-μ

), wherein

Δ

is the Mahalanobis distance,

T is the target characteristic,

μ

is the mean characteristic, and

Σ

<sup>−

1</sup> is the inverse covariance.

33. A system for choosing visual characteristics of a characteristic space within a region, comprising:

a saliency determining circuit, routine or application usable to determine a saliency based on an inverse covariance of the region and a difference between a target characteristic and a mean characteristic of the region;

a comparing circuit, routine or application usable to compare the saliency against at least one acceptance criterion; and

a characteristics adjusting circuit, routine or application usable to adjust the target characteristic if the comparing circuit, routine or application determines that the saliency does not satisfy the at least one acceptance criterion.

34. The system of claim 33, wherein the saliency determining circuit, routine or application includes:

a mean determining circuit, routine or application useable to determine a mean characteristic of the region;

a covariance determining circuit, routine or application usable to determine a covariance of the region;

an covariance inverting circuit, routine or application usable to invert the covariance to produce an inverse covariance;

a characteristic estimating circuit, routine or application usable to estimate a target characteristic;

a differencing circuit, routine or application usable to determine a difference between the mean characteristic and the target characteristic;

a transposing circuit, routine or application usable to transpose the difference to produce a difference transpose; and

a product determining circuit, routine or application usable to determine a product of the difference transpose multiplied by the inverse covariance multiplied by the difference.

35. The system of claim 33, wherein the visual characteristics are colors and the characteristic space is a color space.

36. The system of claim 33, wherein the characteristic space has a plurality of coordinates, the mean characteristic is a vector having vector elements, the covariance is a matrix having tensor elements, and the target characteristic is a vector having vector elements.

37. The system of claim 34, wherein the characteristic estimating circuit, routine or application includes a subtracting circuit, routine or application usable to subtract a vector element for a coordinate of the mean characteristic from a midpoint of the coordinate.

38. The system of claim 33, wherein the saliency determining circuit, routine or application includes a Mahalanobis distance determining circuit, routine or application usable to determine a Mahalanobis distance.

39. The system of claim 38, wherein the Mahalanobis distance determining circuit, routine or application determines the Mahalanobis distance Δ

as;

Δ

²=(T-μ

)′

Σ

<sup>−

1</sup>(T-μ

), where;

T is the target characteristic,

μ

is the mean characteristic, and

Σ

<sup>−

1</sup> is the inverse covariance.