MULTI-LEVEL HALFTONING PROVIDING REDUCED ERROR DIFFUSION ARTIFACTS

• US 20090219577A1
• Filed: 02/29/2008
• Published: 09/03/2009
• Est. Priority Date: 02/29/2008
• Status: Active Grant

• Current Assignee: Eastman Kodak Company
• Original Assignee: Eastman Kodak Company
• Inventors: Spaulding, Kevin E., Couwenhoven, Douglas W.
• Granted Patent: US 8,018,623 B2
• Time in Patent Office: Days
• US Class Current: 358/3.30
• CPC Class Codes: H04N 1/40087 Multi-toning, i.e. converting a continuous-tone signal for reproduction with more than two discrete brightnesses or optical densities, e.g. dots of grey and black inks on white paper
• CPC Class Codes: H04N 1/4053 with threshold modulated relative to input image data or vice versa

Abstract

A method for multi-level error diffusion halftoning an input digital image to form an output digital image with reduced error diffusion texture contouring artifacts which includes modifying the input pixel values for a fraction of the input pixels using a periodic dither signal, where the fraction is a function of the texture artifact characteristics of the multi-level error diffusion halftoning process.

Claims

1. A method for multi-level error diffusion halftoning an input digital image having input pixels with corresponding input pixel values and N_i input levels to form an output digital image with reduced error diffusion texture contouring artifacts, such an output digital image having output pixels with corresponding output pixel values and N_o output levels, where 2≦

N_o<

N_i, comprising;

a) determining a fraction of input pixels to be modified using a periodic dither signal in response to the input pixel value of the input pixel, wherein the fraction of input pixels is a function of the texture artifact characteristics that would result from the application of a multi-level error diffusion halftoning process to a uniform image region having the particular input pixel value, the fraction of input pixels being larger for input pixel values that would produce visually objectionable texture artifact characteristics than for input pixel values that would produce visually pleasing texture characteristics;

b) modifying the input pixel values for the determined fraction of input pixels using a periodic dither signal determined by modularly addressing a dither signal array with the row and column address of the input pixel in the input digital image, wherein the dither signal array is designed to produce periodic dither signals having low visibility to a human observer;

c) applying the multi-level error diffusion halftoning process to the adjusted pixel value to determine the output pixel value; and

d) repeating steps a)-d) for multiple input pixels of the input digital image to thereby provide the output digital image.

2. The method according to claim 1 wherein the dither signal array is determined by reducing a visual cost function.

3. The method according to claim 1 wherein the dither signal array is designed to produce periodic dither signals having spatial frequency characteristics similar to those produced by the multi-level error diffusion halftoning process.

4. The method according to claim 1 wherein the dither signal array is a dither matrix storing periodic dither values, and wherein large dither matrix elements are paired with and are arranged to be adjacent to small dither matrix elements in the dither matrix.

5. The method according to claim 1 wherein step b) includes modifying the input pixel values by applying a multi-level dither halftoning process that uses the periodic dither signal, such that the modified input pixels values correspond to the N_o output levels.

6. The method according to claim 1 wherein the multi-level error diffusion halftoning process is a binary error diffusion halftoning process producing output pixel values with two output levels.

7. The method according to claim 1 wherein the texture characteristics correspond to a perceived granularity value for the halftoning pattern produced by multi-level error diffusion halftoning process.

8. The method according to claim 1 wherein the texture characteristics correspond to a texture visibility value for the halftoning pattern produced by multi-level error diffusion halftoning process.

9. The method according to claim 1 wherein the fraction of input pixels to be modified is adjusted to produce an output digital image having more uniform texture characteristics.

10. The method according to claim 1 wherein the multi-level error diffusion halftoning process in step d) includes:

i) applying a quantization operation to the adjusted pixel value to determine the output pixel value;

ii) determining an error signal by computing a difference between the output pixel value and the adjusted pixel value; and

iii) scaling the error signal by a set of error feedback weights and adding the scaled error weights to the input pixel values for nearby input pixels that have yet to be processed.

11. The method according to claim 1 wherein the multi-level error diffusion halftoning process in step d) includes producing an error feedback signal, and wherein the amplitude of the error feedback signal is adjusted responsive to the input pixel value of the input pixel.

12. The method according to claim 11 wherein the amplitude of the error feedback signal is adjusted to be smaller for input pixel values that would produce visually objectionable texture characteristics when the multilevel error diffusion halftoning process is applied to a uniform image region having the particular input pixel value, and larger for input pixel values that would produce visually pleasing texture characteristics when the multi-level error diffusion halftoning process is applied to a uniform image region having the particular input pixel value.

13. The method according to claim 1 wherein the input digital image is a color input digital image having a plurality of color channels and the multi-level error diffusion halftoning process is a color error diffusion process.

14. The method according to claim 13 wherein different dither signal arrays are used for each color channel of the input digital image.

15. The method according to claim 14 wherein the dither signal arrays used for each color channels are jointly optimized.

16. The method according to claim 13 where the fraction of input pixels to be modified are determined independently for each color channel as a function of the input pixel values for the corresponding color channel.

17. The method according to claim 13 where the fraction of input pixels to be modified is determined using a multi-dimensional dither fraction function that is a function of the input pixel values for multiple color channels of the input digital image.

18. The method according to claim 13 wherein the color multi-level error diffusion process is a vector error diffusion process.

19. A method for multi-level halftoning a color input digital image having a plurality of color channels where the method according to claim 1 is applied to at least one color channel of the color input digital image.