Identifying banding in digital images

1. A method for identifying banding in a digital image comprising:

accessing the digital image;

determining by a processor for a particular pixel in the digital image, based on at least two candidate scales that represent sizes of regions that are evaluated for banding, whether at least one portion of the digital image includes one or more bands having a difference in color, the at least one portion including the particular pixel, and wherein determining whether at least one portion includes one or more bands further comprises:

determining, for the at least two candidate scales, a result indicating whether banding is present at the particular pixel; and

determining a most-likely scale of banding for the particular pixel based on the results for the at least two candidate scales.

2. The method of claim 1 wherein determining whether at least one portion includes one or more bands further comprises:

determining, for at least two candidate scales, a first fraction of pixels, in the region of the digital image that is based on the candidate scale being evaluated, having a color value offset by an offset value from a color value of the particular pixel;

determining, for at least two candidate scales, a second fraction of pixels, in the region of the digital image that is based on the candidate scale being evaluated, having a color value equal to the color value of the particular pixel; and

wherein determining the result for the at least two candidate scales is based on one or more of the first fraction and the second fraction for the respective candidate scale.

3. The method of claim 2 wherein determining the result for the at least two candidate scales comprises determining, for at least one of the candidate scales, whether one or more of the first fraction and the second fraction exceeds a threshold value.

4. The method of claim 1 further comprising:

determining that the digital image includes adjacent bands having a difference in color resulting in a contour between the adjacent bands; and

applying an algorithm to at least a portion of the digital image for reducing visibility of the contour, the algorithm being based on a value representing the fraction of pixels in a region of the digital image having a particular color value.

5. The method of claim 4 wherein applying the algorithm comprises dithering pixels in the portion using a dithering algorithm that selects a color value for a pixel in the portion based on the value representing the fraction.

6. The method of claim 1 wherein the most-likely scale is indicative of a width of the one or more bands.

7. The method of claim 6 wherein the digital image includes multiple color components and determining the result for the at least two candidate scales is based on an evaluation of corresponding pixels in the multiple color components.

8. The method of claim 1 wherein the digital image is a single color component.

9. The method of claim 1 wherein the digital image includes quantized pixels, and the bands result from the quantization of the quantized pixels.

10. The method of claim 1 wherein the one or more bands includes two adjacent bands and the difference in color between the two adjacent bands results in a contour between the two adjacent bands.

11. The method of claim 1 wherein determining the results for the at least two candidate scales comprises determining a confidence score associated with the at least two candidate scales; and

determining the most-likely scale is based on confidence scores associated with the at least two candidate scales.

12. The method of claim 11 wherein determining the confidence score comprises using a two-part equation including a part indicating probability of banding and another part indicating visual significance of banding.

13. The method of claim 11 wherein determining the most-likely scale comprises selecting a scale having the largest confidence score that satisfies a thresholding test.

14. The method of claim 11 wherein the confidence score for at least one of the at least two candidate scales is determined by:

determining a region of pixels in the digital image based on the at least one candidate scale; and

determining a confidence score for the at least one candidate scale based on color values of pixels in the determined region.

15. The method of claim 14 wherein determining a confidence score comprises determining a fraction of pixels having at least one selected color value in the determined region.

16. The method of claim 15 wherein the at least one selected color value is determined based on one or more selected offset values from a color value of at least one pixel in the determined region.

17. The method of claim 1 wherein the digital image corresponds to at least one component of a multi-component image.

18. The method of claim 1 wherein the candidate scales indicate width of bands.

19. A method for identifying banding in a digital image comprising:

accessing the digital image;

assessing by a processor at least a portion of the digital image for the existence of one or more bands having a difference in color, wherein assessing includes determining a fraction of pixels in the portion having a color value offset by an offset value from a color value of a particular pixel in the portion.

20. The method of claim 19 wherein assessing includes determining that there is at least one band in the portion if the determined fraction exceeds a threshold.

21. The method of claim 19 further comprising determining a second fraction of pixels in the image portion having a color value offset by a second selected value from the color value of the at least one pixel in the image portion.

22. The method of claim 21 wherein assessing comprises: comparing the determined fraction to a first threshold;

comparing the determined second fraction to a second threshold; and

determining whether there is at least one band in the portion based on (1) a result of comparing the determined fraction to the first threshold and (2) a result of comparing the determined second fraction to a second threshold.

23. The method of claim 22 wherein determining whether there is at least one band comprises determining that there is at least one band in the portion if (1) the determined fraction exceeds the first threshold and (2) the determined second fraction exceeds the second threshold.

24. The method of claim 19 wherein the selected offset value is one.

25. The method of claim 19 wherein the one or more bands includes two adjacent bands and the difference in color between the two adjacent bands results in a contour between the two adjacent bands.

26. An apparatus for identifying banding in a digital image comprising:

means for accessing the digital image;

means for determining for a particular pixel in the digital image, based on at least two candidate scales that represent sizes of regions that are evaluated for banding, whether at least one portion of the digital image includes one or more bands having a difference in color, the at least one portion including the particular pixel, and wherein the means for determining whether at least one portion includes one or more bands further comprises:

means for determining, for the at least two candidate scales, a result indicating whether banding is present at the particular pixel; and

means for determining a most-likely scale of banding for the particular pixel based on the results for the at least two candidate scales.

27. The apparatus of claim 26 further comprising a storage unit for storing the digital image.

28. The apparatus of claim 26 wherein the apparatus is part of an encoder.

29. The apparatus of claim 26 wherein the apparatus is part of a decoder.

30. The apparatus of claim 26 wherein the apparatus is part of a filter.

31. A non-transitory processor-readable medium having one or more executable instructions stored thereon, which when executed by a digital processing system cause the digital processing system to perform a method for identifying banding in a digital image, the method comprising:

accessing the digital image;

determining for a particular pixel in the digital image, based on at least two candidate scales that represent sizes of regions that are evaluated for banding, whether at least one portion of the digital image includes one or more bands having a difference in color, the at least one portion including the particular pixel, and wherein the means for determining whether at least one portion includes one or more bands further comprises:

determining, for the at least two candidate scales, a result indicating whether banding is present at the particular pixel; and

determining a most-likely scale of banding for the particular pixel based on the results for the at least two candidate scales.