Banding artifact detection in digital video content
A method and system for identifying and determining banding artifacts in digital video content composed of a sequence of moving video pictures includes creating a mask image corresponding to a picture from said sequence of moving video pictures based on global gradient changes to detect potential areas containing banding artifacts. The values of the mask image are scaled thereby making banding artifact detection possible using gradient operators. The banding artifacts are then identified/detected based on the local gradients.
1. A method of processing digital video content comprising the steps of:
dividing at least one area of a video picture from said digital video content into blocks;
computing an average pixel value for each of said blocks;
forming a DC plane comprising a plurality of the average pixel values;
computing a gradient for each of the plurality of said average pixel values of said DC plane;
computing at least a mean and variance of the gradients to determine whether said at least one area is at least one of a texture area, a smooth area, and a slow change area by comparing said at least a mean and variance to a respective threshold value;
excluding said at least one area from banding artifact detection if said at least one area is categorized as the texture area; and
detecting whether said at least one area of said video picture has a banding artifact if said at least one area is categorized as the smooth area or the slow change area.
2. The method of claim 1 , comprising the step of scaling said at least one area if said at least one area is categorized as the smooth area or the slow change area.
3. The method of claim 1 , comprising the step of:
categorizing said at least one area of said video picture as having said banding artifact based on local gradients.
4. The method of claim 1 , and said method comprises the additional steps of:
marking said at least one area as the texture area when the at least a mean and variance is greater than said respective threshold; and
marking said at least one area as at least one of the smooth area and the slow change area when one of the at least a mean and variance value is less than said respective threshold.
5. The method of claim 1 , wherein said excluding step comprises:
replacing said at least one area with a constant value if said at least one area is categorized as the texture area.
6. The method of claim 1 , wherein said detecting step further comprises:
applying a pixel based gradient computation to determine a local gradient, wherein the detecting step is responsive to the local gradient.
7. A program storage device having an application program tangibly embodied thereon, the application program including instructions for performing at least the following on a video picture from a sequence of video pictures:
dividing at least one area of said video picture from said sequence of video pictures into blocks;
computing an average pixel value for each of said blocks;
forming a DC plane comprising a plurality of the average pixel values;
computing a gradient for each of the plurality of said average pixel values of said DC plane;
computing at least a mean and variance of the gradients;
comparing the at least a mean and variance of the gradients to a respective threshold;
marking said at least one area as a texture area when the at least a mean and variance is greater than said respective threshold;
marking said at least one area as at least one of a smooth area and a slow change area when the at least a mean and variance value is less than said respective threshold;
excluding said at least one area from banding artifact detection if said at least one area is categorized as the texture area; and
detecting banding artifacts in said at least one area if said at least one area is categorized as the smooth area or the slow change area.
8. An apparatus for processing digital video content, comprising:
a banding artifact detector configured to
divide at least one area of a video picture from said digital video content into blocks;
compute an average pixel value for each of said blocks;
form a DC plane comprising a plurality of the average pixel values;
compute a gradient for each of the plurality of said average pixel values of said DC plane;
compute at least a mean and variance of the gradients as to determine whether said at least one area is at least one of a texture area, a smooth area, and a slow change area by comparing said at least a mean and variance to a respective threshold value;
exclude said at least one area from banding artifact detection if said at least one area is categorized as the texture area; and
detect whether said at least one area has a banding artifact if said at least one area is categorized as the smooth area or the slow change area.
9. The apparatus of claim 8 , wherein the banding artifact detector further performs:
marking said at least one area as the texture area when the at least a mean and variance is greater than said respective threshold; and
marking said at least one area as at least one of the smooth area and the slow change area when one of the at least a mean and variance value is less than said respective threshold.