Image rendering system and method for generating stochastic threshold arrays for use therewith

• US 5,745,660 A
• Filed: 04/26/1995
• Issued: 04/28/1998
• Est. Priority Date: 04/26/1995
• Status: Expired due to Term

• Current Assignee: Intellectual Ventures I LLC (Intellectual Ventures LLC)
• Original Assignee: Polaroid Corporation (PLR IP Holdings LLC)
• Inventors: Kolpatzik, Bernd W., Thornton, Jay E.
• Primary Examiner(s): Coles, Sr., Edward L.
• Assistant Examiner(s): Harris, Tia M.
• Application Number:
  US08/427,880
• Time in Patent Office: 1,098 Days
• Field of Search: 395/108, 358/456, 358/457, 358/465, 358/429, 358/448, 358/455, 358/458, 358/459, 358/466, 382/270, 382/272
• US Class Current: 358/3.19
• CPC Class Codes: H04N 1/4051 producing a dispersed dots halftone pattern, the dots having substantially the same size different sizes H04N1/4057

Abstract

An image rendering system, for generating discrete tone image data representing a continuous-tone image for rendering by a rendering device, the image rendering system includes a stochastic threshold array generating arrangement and an image data processing section. The stochastic threshold array generating arrangement generates a plurality of stochastic threshold arrays from a random seed pattern that have been subsequently optimized to reduce image artifacts in the rendering of tones. Each pattern is optimized to reflect (i) a selected trade-off between grain and possible mottle in the rendered image, (ii) distortions introduced by the rendering device in rendering an image defined by the pattern and an observer in viewing the image, and (iii) boundary effects so that the patterns, when tiled together, have optimal visual characteristics proximate their respective edges. The image data processing section processes the continuous tone image data in relation to the stochastic threshold arrays generated by the stochastic threshold array generating arrangement. The image data processing section may be used to either randomly tile, or tile in an ordered way, the stochastic threshold arrays over the continuous-tone image data and generates the discrete tone image data in response to correspondingly positioned pixels of the continuous-tone image data and threshold values of the stochastic threshold arrays.

Claims

1. An image rendering system for generating discrete tone image data of an image for representing continuous tone image data to be rendered by a rendering device, the image rendering system comprising:

A. a stochastic threshold array library for storing a plurality of stochastic threshold arrays generated from individual seed patterns each corresponding to a predetermined tone of the image, each said seed pattern optimized to reflect corresponding predetermined grain/mottle trade-off conditions and optimal visual tiling characteristics when said seed pattern is tiled with itself; and

B. an image data processing section for processing the continuous tone image data in relation to said stochastic threshold arrays, and for retrieving one said stochastic threshold array for one of the predetermined grain/mottle trade-off conditions, as selected by an operator, from said stochastic threshold array library, tiling the retrieved stochastic threshold array over the continuous-tone image data and generating the discrete tone image data in response to correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold array.

16. The image rendering system of claim 1 in which each said stochastic threshold array is generated according to the steps of:

A. generating an optimized stochastic pattern from a seed stochastic pattern according to the steps of;

i. generating an initial stochastic pattern; and

ii. optimizing the initial stochastic pattern by, in a series of iterations,

a. generating an error map reflecting differences between a stochastic pattern provided thereto and a rendering device filtered stochastic pattern that represents processing performed by the rendering device in rendering an image defined by the stochastic pattern and a neighborhood filtered stochastic pattern representing psychovisual processing as performed by an observer in viewing the image, respective contributions of the rendering device filtered stochastic pattern and the neighborhood filtered stochastic pattern to the error map being determined by respective complementary weighting factors, differing complementary weighting factors being used during optimizing for the stochastic threshold arrays thereby to provide differing grain/mottle tradeoff conditions;

b. generating an updated stochastic pattern in relation to the error map generated during the iteration, the update stochastic pattern being used as the stochastic pattern used in generating the error map during a subsequent iteration; and

B. generating the stochastic threshold arrays from the optimized stochastic patterns.

17. The image rendering system of claim 16 in which each stochastic pattern comprises a plurality of pixels, and in which the error map generation step comprises the steps of:

A. generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. generating a composite error map which provides a grain/mottle tradeoff in response to both the rendering device error map and the neighborhood error map using the complementary weighting values to determine respective contributions of the rendering device error map and the neighborhood error map to the composite error map.

18. The image rendering system of claim 17 in which the rendering device error map generating step includes the steps of:

A. performing a rendering device filtering operation in connection with the stochastic pattern to generate a rendering device filtered stochastic pattern, the rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering large clusters of pixels comprising the stochastic pattern;

B. performing a contrast sensitivity filtering operation in connection with the rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern, the contrast sensitivity filtered stochastic pattern reflecting the psychovisual processing performed by the observer in connection with the observing the rendering device filtered stochastic pattern; and

C. generating the rendering device error map identifying differences between the stochastic pattern and the contrast sensitivity filtered stochastic pattern in connection with large clusters of pixels.

19. The image rendering system of claim 18 in which, during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern, the rendering device filtering operation and the contrast sensitivity filtering operation are both performed in accordance with a circular convolution filtering methodology.

20. The image rendering system of claim 17 in which the composite error map generation step comprises the steps of:

A. performing a neighborhood filtering operation in connection with the stochastic pattern to generate a neighborhood filtered stochastic pattern, the neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering small clusters of pixels comprising the stochastic pattern; and

B. generating the neighborhood error map identifying differences between the stochastic pattern and the neighborhood filtered stochastic pattern in connection with small clusters of pixels.

21. The image rendering system of claim 20 in which, during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern, the neighborhood filtering operation is performed in accordance with a circular convolution filtering methodology.

22. The image rendering system of claim 20 in which the neighborhood error map is generated from the stochastic pattern using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function, the predetermined Gaussian function having an parameter that may be varied to vary the filtered cluster size.

23. The image rendering system of claim 22 in which the neighborhood error map is generated from the stochastic pattern using a filter applied to each pixel of the stochastic pattern of the form ##EQU3## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

2. An image rendering system for generating discrete tone image data representing continuous-tone image data for rendering by a rendering device, the image rendering system comprising:

A. a stochastic threshold array library for storing a plurality of stochastic threshold arrays generated from individual seed patterns optimized to have optimal visual tiling characteristics when randomly tiled together; and

B. an image data processing section for processing the continuous tone image data in relation to the stochastic threshold arrays, randomly retrieving the stochastic threshold arrays from said stochastic threshold array library, tiling said retrieved stochastic threshold arrays randomly over the continuous-tone image data and generating the discrete tone image data in response to both correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold arrays.

3. The image rendering system of claim 2 in which said stochastic threshold arrays are generated according to the steps of:

A. generating a plurality of optimized stochastic patterns each from a respective seed stochastic pattern, at least some of the optimized stochastic patterns each being generated from said respective seed stochastic pattern and a previously-generated optimized stochastic pattern to thereby provide that the optimized stochastic pattern so generated, when tiled with others of the optimized stochastic patterns, will have optimal visual tiling characteristics proximate their respective edges; and

B. generating each stochastic threshold array using one of said optimized stochastic patterns.

4. The image rendering system of claim 3 in which said optimized stochastic pattern generating step comprises the steps of:

A. generating an initial stochastic pattern, the initial stochastic pattern for the first iteration comprising said respective seed stochastic pattern, and at least some initial stochastic patterns for use in generating optimized stochastic patterns after the first optimized stochastic pattern including said respective seed stochastic pattern surrounded by copies of at least one optimized stochastic pattern generated during a previous iteration; and

B. optimizing the initial stochastic pattern.

5. The image rendering system of claim 4 in which the stochastic pattern optimizing step comprises the steps of iteratively:

A. generating an error map reflecting differences between a stochastic pattern provided thereto and processing performed by the rendering device in rendering an image defined by the stochastic pattern and psychovisual processing as performed by an observer in viewing the image;

B. generating an updated stochastic pattern in relation to the error map generated during the iteration, the updated stochastic pattern being used as the stochastic pattern used in generating the error map during a subsequent iteration.

6. The image rendering system of claim 5 in which each stochastic pattern comprises pixels represented by two complementary colors, the updated stochastic pattern update pattern generating step comprising the step of alternating the colors of selected numbers of pixels of each color to their respective complementary colors, the number being selectively reduced for successive iterations.

7. The image rendering system of claim 5 in which each stochastic pattern comprises a plurality of pixels, and in which the error map generation step comprises the steps of:

A. generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. generating a composite error map which provides a grain/mottle tradeoff in response to both the rendering device error map and the neighborhood error map.

8. The image rendering system of claim 7 in which the rendering device error map generating step includes the steps of:

A. performing a rendering device filtering operation in connection with the stochastic pattern to generate a rendering device filtered stochastic pattern, the rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering large clusters of pixels comprising the stochastic pattern;

B. performing a contrast sensitivity filtering operation in connection with the rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern, the contrast sensitivity filtered stochastic pattern reflecting the psychovisual processing performed by the observer in connection with the observing the rendering device filtered stochastic pattern; and

C. generating the rendering device error map identifying differences between the stochastic pattern and the contrast sensitivity filtered stochastic pattern in connection with large clusters of pixels.

9. The image rendering system of claim 8 in which, during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern, the rendering device filtering operation and the contrast sensitivity filtering operation are both performed in accordance with a circular convolution filtering methodology.

10. The image rendering system of claim 7 in which the composite error map generation step comprises the steps of:

A. performing a neighborhood filtering operation in connection with the stochastic pattern to generate a neighborhood filtered stochastic pattern, the neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering small clusters of pixels comprising the stochastic pattern; and

B. generating the neighborhood error map identifying differences between the stochastic pattern and the neighborhood filtered stochastic pattern in connection with small clusters of pixels.

11. The image rendering system of claim 10 in which, during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern, the neighborhood filtering operation is performed in accordance with a circular convolution filtering methodology.

12. The image rendering system of claim 10 in which the neighborhood error map is generated from the stochastic pattern using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function, the predetermined Gaussian function having an parameter that may be varied to vary the filtered cluster size.

13. The image rendering system of claim 12 in which the neighborhood error map is generated from the stochastic pattern using a filter applied to each pixel of the stochastic pattern of the form ##EQU2## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

14. The image rendering system of claim 7 in which the composite error map generation step includes the steps of:

A. generating a weighted rendering device error map from the rendering device error map and a weighted neighborhood error map from the neighborhood error map both from a weighting value; and

B. generating said combined error map from the weighted rendering device error map and the weighted neighborhood error map.

15. The image rendering system of claim 14 in which the weighting value is selected by an operator.

24. A stochastic threshold array generating subsystem for generating a plurality of stochastic threshold arrays for use in connection with generation of output image data from continuous tone image data for rendering by a rendering device, said stochastic threshold array generating subsystem comprising:

A. a stochastic pattern generator for generating a plurality of optimized stochastic patterns each from a respective seed stochastic pattern, at least one of said optimized stochastic patterns being generated in response to both said respective seed stochastic pattern and a previously generated stochastic pattern, said at least one pattern, when tiled with others of said optimized stochastic patterns, having optimal visual tiling characteristics proximate to respective edges of said tiled patterns; and

B. a stochastic threshold array generator for generating the stochastic threshold arrays from said optimized stochastic patterns generated by said stochastic pattern generator.

25. The stochastic threshold array generating subsystem of claim 24 in which said stochastic pattern generator comprises:

A. an initial stochastic pattern generator for generating initial stochastic patterns, a first initial stochastic pattern for a first iteration comprising said seed stochastic pattern, and at least one of said initial stochastic patterns for use in generating said optimized stochastic patterns after a first optimized stochastic pattern including said seed stochastic pattern surrounded by copies of at least one of said optimized stochastic patterns generated during a previous iteration; and

B. a stochastic pattern optimizer for receiving said initial stochastic patterns and processing said initial stochastic patterns to generate respective said optimized stochastic patterns.

26. The stochastic threshold array generating subsystem of claim 25 in which said stochastic pattern optimizer comprises:

A. a first error map generator for generating a first error map reflecting differences between said initial stochastic patterns provided thereto and processing performed by the rendering device in rendering an image defined by said optimized stochastic patterns and psychovisual processing as performed by an observer in viewing the image;

B. stochastic pattern update generator for generating updated said optimized stochastic patterns in relation to said first error map;

C. an iteration control for controlling said first error map generator and said stochastic pattern update generator through a series of iterations to generate said optimized stochastic patterns, wherein in said first iteration said iteration control enables said first error map generator to operate in connection with said initial stochastic patterns and in subsequent iterations said iteration control enables said first error map generator to operate in connection with said updated optimized stochastic patterns.

27. The stochastic threshold array generating subsystem of claim 26 in which each said stochastic pattern comprises pixels represented by two complementary colors, said stochastic pattern update generator updating said optimized stochastic patterns by alternating colors of selected numbers of pixels of each color to respective complementary colors, a number being selected for each iteration by the iteration control, the iteration control reducing the number for successive iterations.

28. The stochastic threshold array generating subsystem of claim 26 in which each said optimized stochastic pattern comprises a plurality of pixels, and in which said error map generator comprises:

A. a rendering device model error map generator for generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of said image represented by each said stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. a neighborhood error map generator for generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of said image represented by said optimized stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. a composite error map generator for generating a composite error map which provides a grain/mottle tradeoff in response to both said rendering device error map and said neighborhood error map.

29. The stochastic threshold array generating subsystem of claim 28 in which the rendering device error map generator includes:

A. a rendering device filter for performing a rendering device filtering operation in connection with said stochastic patterns provided to said first error map generator to generate a rendering device filtered stochastic pattern, said rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering said large clusters of pixels;

B. a contrast sensitivity filter for performing a contrast sensitivity filtering operation in connection with said rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern reflecting the psychovisual processing performed by the observer in connection with observing said rendering device filtered stochastic pattern; and

C. a second error map generator for generating rendering device error map identifying differences between said stochastic patterns provided to said first error map generator and said contrast sensitivity filtered stochastic pattern in connection with said large clusters of pixels.

30. The stochastic threshold array generating subsystem of claim 29 in which, during processing of said initial stochastic patterns and updated stochastic patterns while generating said first optimized stochastic pattern, said rendering device filter performs the rendering device filtering operation and said contrast sensitivity filter performs the contrast sensitivity filtering operation both in accordance with a circular convolution filtering methodology.

31. The stochastic threshold array generating subsystem of claim 28 in which said composite error map generator includes:

A. a neighborhood filter for performing a neighborhood filtering operation in connection with said stochastic patterns provided to said second error map generator to generate a neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering said small clusters of pixels;

B. a third error map generator for generating the neighborhood error map identifying differences between said stochastic patterns as provided to said second error map generator and said neighborhood filtered stochastic pattern in connection with said small clusters of pixels.

32. The stochastic threshold array generating subsystem of claim 31 in which, during processing of said initial stochastic patterns and updated stochastic patterns while generating said first optimized stochastic pattern, said neighborhood filter performs the neighborhood filtering operation in accordance with a circular convolution filtering methodology.

33. The stochastic threshold array generating subsystem of claim 31 in which said third error map generator generates said neighborhood error map from said optimized stochastic patterns using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function having a parameter that varies the filtered cluster size.

34. The stochastic threshold array generating subsystem of claim 33 in which said third error map generator generates said neighborhood error map from said optimized stochastic patterns using a filter applied to each pixel of said optimized stochastic patterns of the form ##EQU4## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

35. The stochastic threshold array generating subsystem of claim 28 in which said composite error map generator includes:

A. weighted error generating means for generating a weighted rendering device error map from said rendering device error map and a weighted neighborhood error map from said neighborhood error map both in response to a weighting value; and

B. a combiner for generating said combined error map from said weighted rendering device error map and said weighted neighborhood error map.

36. The stochastic threshold array generating subsystem of claim 35 in which said weighting value is selected by an operator.

37. A stochastic threshold array generating subsystem for generating a plurality of stochastic threshold arrays for use in connection with generation of output image data from continuous tone image data for rendering by a rendering device, said stochastic threshold array generating subsystem comprising:

A. an optimized stochastic pattern generator for generating an optimized stochastic pattern from a seed stochastic pattern comprising;

i. an initial stochastic pattern generator for generating an initial stochastic pattern; and

ii. a stochastic pattern optimizer for optimizing the initial stochastic pattern comprising;

a. an error map generator for generating an error map reflecting differences between a stochastic pattern provided thereto and a rendering device filtered stochastic pattern that represents processing performed by the rendering device in rendering an image followed by psychovisual processing as performed by an observer in viewing the image and a neighborhood filtered stochastic pattern which identifies and penalizes features associated with mottle, respective contributions of the rendering device filtered stochastic pattern and the neighborhood filtered stochastic pattern to the error map being determined by respective complementary weighting factors, differing weighting factors being used during optimizing for the stochastic threshold arrays thereby to provide differing grain/mottle tradeoff conditions;

b. an updated stochastic pattern generator for generating an updated stochastic pattern in relation to the error map generated during the iteration; and

c. an iteration control for controlling the error map generator and the updated stochastic pattern generator through a series of iterations, the iteration control enabling the error map generator to use the updated stochastic pattern generated during an iteration as the stochastic pattern used in generating the error map during a subsequent iteration;

B. generating the stochastic threshold arrays from the optimized stochastic patterns.

38. The stochastic threshold array generating subsystem of claim 37 in which each stochastic pattern comprises a plurality of pixels, and in which the error map generator comprises:

A. a rendering device error map generator generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. a neighborhood error map generator for generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. a composite error map generator for generating a composite error map which provides a grain/mottle tradeoff in response to both the rendering device error map and the neighborhood error map using the complementary weighting values to determine respective contributions of the rendering device error map and the neighborhood error map to the composite error map.

39. The stochastic threshold array generating subsystem of claim 38 in which the rendering device error map generator comprises:

A. a rendering device filter for performing a rendering device filtering operation in connection with the stochastic pattern to generate a rendering device filtered stochastic pattern, the rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering large clusters of pixels comprising the stochastic pattern;

B. a contrast sensitivity filter for performing a contrast sensitivity filtering operation in connection with the rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern, the contrast sensitivity filtered stochastic pattern reflecting the psychovisual processing performed by the observer in connection with the observing the rendering device filtered stochastic pattern; and

C. a composite rendering device error map generator for generating the rendering device error map identifying differences between the stochastic pattern and the contrast sensitivity filtered stochastic pattern in connection with large clusters of pixels.

40. The stochastic threshold array generating subsystem of claim 39 which, during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern, the rendering device filtering operation and the contrast sensitivity filtering operation are both performed in accordance with a circular convolution filtering methodology.

41. The stochastic threshold array generating subsystem of claim 38 in which the neighborhood error map generator comprises:

A. a neighborhood filter for performing a neighborhood filtering operation in connection with the stochastic pattern to generate a neighborhood filtered stochastic pattern, the neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering small clusters of pixels comprising the stochastic pattern; and

B. an error map generator for generating the neighborhood error map identifying differences between the stochastic pattern and the neighborhood filtered stochastic pattern in connection with small clusters of pixels.

42. The stochastic threshold array generating subsystem of claim 41 in which the neighborhood filter performs the neighborhood filtering operation in accordance with a circular convolution filtering methodology during processing of the initial stochastic pattern and updated stochastic patterns while generating the first optimized stochastic pattern.

43. The stochastic threshold array generating subsystem of claim 41 in which the neighborhood filter performs the neighborhood filtering operation using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function, the predetermined Gaussian function having an parameter that may be varied to vary the filtered cluster size.

44. The stochastic threshold array generating subsystem of claim 43 in which the neighborhood filter performs the neighborhood filtering operation using a filter applied to each pixel of the stochastic pattern of the form ##EQU5## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

45. An error map generator for generating an error map for use in generating updated stochastic patterns in connection with a stochastic pattern provided to the error map generator, the stochastic pattern to be used in connection with generation of a stochastic threshold array to be used in generating output image data from continuous-tone image data to be rendered by a rendering device, each stochastic pattern comprising a plurality of pixels, the error map generator comprising:

A. means for generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. a neighborhood error map generator for generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of said image represented by the stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. a composite error map generator for generating a composite error map which provides a grain/mottle tradeoff in response to both the rendering device error map and the neighborhood error map.

46. The error map generator of claim 45 in which said generating means comprises:

A. a rendering device filter for performing a rendering device filtering operation in connection with the stochastic pattern provided to the error map generator to generate a rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering said large clusters of pixels;

B. a contrast sensitivity filter for performing a contrast sensitivity filtering operation in connection with the rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern reflecting psychovisual processing performed by an observer in connection with observing the rendering device filtered stochastic pattern; and

C. a first generator for generating the rendering device error map by identifying differences between the stochastic pattern as provided to the error map generator and the contrast sensitivity filtered stochastic pattern in connection with said large clusters of pixels.

47. The error map generator of claim 45 in which the composite error map generator includes:

A. a neighborhood filter for performing a neighborhood filtering operation in connection with the stochastic pattern provided to the error map generator to generate a neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering said small clusters of pixels;

B. a second generator for generating the neighborhood error map by identifying differences between the stochastic pattern as provided to the error map generator and the neighborhood filtered stochastic pattern in connection with said small clusters of pixels.

48. The error map generator of claim 47 in which the neighborhood filter performs the neighborhood filtering operation using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function having a parameter that varies the filtered cluster size.

49. The error map generator of claim 48 in which the neighborhood filter performs the neighborhood filtering operation using a filter applied to each pixel of the stochastic pattern of the form ##EQU6## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

50. The error map generator of claim 45 in which the composite error map generator includes:

A. weighted error generating means for generating a weighted rendering device error map from the rendering device error map and a weighted neighborhood error map from the neighborhood error map both from a weighting value; and

B. a combiner for generating said combined error map from the weighted rendering device error map and the weighted neighborhood error map.

51. The error map generator of claim 50 in which the weighting value is selected by an operator.

52. A method of generating an error map for use in generating updated stochastic patterns to be used in connection with generation of a stochastic threshold array to be used in generating output image data from continuous-tone image data to be rendered by a rendering device, each stochastic pattern comprising a plurality of pixels, the method comprising the steps of:

A. generating a rendering device error map which minimizes grain by identifying errors in connection with processing by the rendering device of an image represented by the stochastic pattern provided thereto in connection with rendering of large clusters of pixels;

B. generating a neighborhood error map which minimizes mottle by identifying errors in connection with processing by the rendering device of said image represented by the stochastic pattern provided thereto in connection with rendering of small clusters of pixels; and

C. generating a composite error map which provides a grain/mottle tradeoff in response to both the rendering device error map and the neighborhood error map.

53. The method of claim 52 in which the rendering device error map generating step includes the steps of:

A. performing a rendering device filtering operation in connection with the stochastic pattern to generate a rendering device filtered stochastic pattern reflecting processing performed by the rendering device in rendering said large clusters of pixels;

B. performing a contrast sensitivity filtering operation in connection with the rendering device filtered stochastic pattern to generate a contrast sensitivity filtered stochastic pattern reflecting psychovisual processing performed by an observer in connection with observing the rendering device filtered stochastic pattern; and

C. generating said rendering device error map by identifying differences between the stochastic pattern and contrast sensitivity filtered stochastic pattern in connection with said large clusters of pixels.

54. The method of claim 53 in which the composite error map generation step includes the steps of:

A. performing a neighborhood filtering operation in connection with generating a neighborhood filtered stochastic pattern, said neighborhood filtered stochastic pattern reflecting processing performed by the rendering device in rendering said small clusters of pixels;

B. generating the neighborhood error map identifying differences between the stochastic pattern as provided to the error map generator and the neighborhood filtered stochastic pattern in connection with small clusters of pixels.

55. The method of claim 54 in which the neighborhood filtering operation is performed using a filter in which each pixel is generated according to the form of a constant value minus a predetermined Gaussian function having a parameter that varies the filtered cluster size.

56. The method of claim 55 in which the neighborhood filtering operation is performed using a filter applied to each pixel of the stochastic pattern of the form ##EQU7## where "x" and "y" are horizontal and vertical axes, respectively, of the stochastic pattern, "x0" and "y0" are coordinates along the respective "x" and "y" axes of the pixel being filtered, "k" and "F" are constants having selected values, with F being variable to vary the filtered cluster size, and "δ

" is a delta function corresponding to the value "1" where (x,y)=(x0,y0) and zero elsewhere.

57. The method of claim 52 in which the composite error map generation step includes the steps of:

A. generating, in response to a weighting value, both a weighted rendering device error map from the rendering device error map and a weighted neighborhood error map from the neighborhood error map; and

B. generating said combined error map from both said weighted rendering device error map and said weighted neighborhood error map.

58. The method of claim 57 in which the weighting value is selected by an operator.

59. A method of generating discrete tone image data from continuous-tone image data for rendering by a rendering device, comprising the steps of:

A. providing a plurality of stochastic threshold arrays generated from individually generated stochastic patterns, each stochastic pattern that is used in an image being optimized to reflect processing performed by the rendering device in rendering said image defined by the stochastic pattern, and being optimized by psychovisual processing performed by an observer in observing the image, and further being optimized so that said stochastic patterns, when tiled together, have optimal visual tiling characteristics proximate to respective edges; and

B. processing the continuous tone image data in relation to ones of said stochastic threshold arrays randomly selected and tiled over the continuous-tone image data, the discrete tone image data being generated from correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold arrays.

60. A method of generating discrete tone image data from continuous-tone image data for rendering by a rendering device, comprising the steps of:

A. providing a plurality of stochastic threshold arrays generated from individually generated stochastic patterns, each stochastic pattern being optimized to reflect corresponding predetermined grain/mottle trade-off conditions and further being optimized so that each pattern, any of which corresponds to a tone from a stochastic threshold array, when tiled with itself has an optimal visual tiling characteristic proximate respective edges; and

B. processing the continuous tone image data in relation to the stochastic threshold arrays by retrieving the stochastic threshold array for one of the predetermined grain/mottle trade-off conditions, as selected by an operator, from a stochastic threshold array library, tiling the retrieved stochastic threshold array over the continuous-tone image data and generating the discrete tone image data from correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold array.

61. A method of generating a plurality of stochastic threshold arrays for use in connection with generation of output image data from continuous tone image data for rendering by a rendering device, comprising the steps of:

A. generating a plurality of optimized stochastic patterns each from a respective seed stochastic pattern, at least one of the optimized stochastic patterns being generated from said respective seed stochastic pattern and a previously-generated optimized stochastic pattern to thereby provide that the optimized stochastic pattern so generated, when tiled with others of the optimized stochastic patterns, will have optimal visual tiling characteristics proximate their respective edges; and

B. generating the stochastic threshold arrays from the optimized stochastic patterns.

62. A stochastic threshold array library comprising a plurality of stochastic threshold arrays for use in connection with generation of output image data from continuous tone image data for rendering by a rendering device, the stochastic threshold array library being generated in accordance with the steps of:

A. generating a plurality of optimized stochastic patterns each from a respective seed stochastic pattern, at least some of the optimized stochastic patterns each being generated from a said respective seed stochastic pattern and a previously-generated optimized stochastic pattern to thereby provide that the optimized stochastic pattern so generated, when tiled with others of the optimized stochastic patterns, will have optimal visual tiling characteristics proximate their respective edges; and

B. generating the stochastic threshold arrays from the optimized stochastic patterns.

63. A discrete tone image generated from continuous-tone image data in accordance with the method comprising the steps of:

A. processing the continuous tone image data in relation to randomly selected stochastic threshold arrays, the randomly selected stochastic threshold arrays being generated from individually-generated stochastic patterns, each stochastic pattern being optimized to reflect processing performed by a selected rendering device in rendering an image defined by the stochastic pattern and psychovisual processing performed by an observer in observing the image, and further being optimized so that the stochastic patterns, when tiled together, have optimal visual tiling characteristics proximate their respective edges, the stochastic threshold arrays being randomly tiled over the continuous-tone image data and the discrete tone image data being generated from correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold arrays; and

B. rendering the output image data using the rendering device.

64. A discrete tone image generated from continuous-tone image data in accordance with the method comprising the steps of:

A. processing the continuous tone image data in relation to randomly selected stochastic threshold arrays generated from individually generated stochastic patterns, each stochastic pattern being optimized to reflect corresponding predetermined grain/mottle trade-off conditions and further being optimized so that when tiled each said stochastic pattern, corresponding to a tone from one of said stochastic threshold arrays, has an optimal visual tiling characteristic proximate respective edges, said stochastic threshold arrays being randomly tiled over the continuous-tone image data, discrete tone image data being generated in response to both correspondingly positioned pixels of the continuous-tone image data and threshold values of the tiled stochastic threshold arrays; and

B. rendering the discrete tone image from said discrete tone image data.