Method and apparatus for encoding a digital motion video signal using pyramid specific filtering
AssigneeIntel Corporation
InventorStuart J. Golin
Life of this patent
Estimating the term…
Abstract
An image having at least one high frequency feature is encoded by (a) filtering the image to produce a filtered image representing the image minus the at least one high frequency feature; (b) resolving the filtered image into at least one lower level of resolution; (c) encoding a correction image at the at least one lower level of resolution; (d) producing an expanded result image having the same the level of resolution of the image; and (e) encoding at the level of resolution the image the difference between the image and the expanded result image.
Claims14
A method for encoding an image having at least one high frequency feature comprising the steps of:
- (a) filtering said image with a non-linear filter targeted at said at least one high resolution feature to produce a filtered image, wherein said at least one high resolution feature is selected from a group comprising lines and points, and wherein said filtering is targeted at said at least one high resolution feature and not at edges;
- (b) low pass filtering and subsampling said filtered image into at least one lower level of resolution;
- (c) encoding a correction image at said at least one lower level of resolution;
- (d) producing an expanded result image from said correction image, said expanded result image having the same level of resolution as said image;
- and
- (e) encoding at the level of resolution of said image the difference between said image and said expanded result image.
A method for filtering an image having at least one high frequency feature comprising the steps of:
- (a) filtering said image with a non-linear filter targeted at said at least one high resolution feature to produce a filtered image, wherein said at least one high resolution feature is selected from a group comprising lines and points, and wherein said filtering is targeted at said at least one high resolution feature and not at edges;
- (b) resolving said filtered image into at least one lower level of resolution;
- and
- (c) analyzing the difference of said image and the expansion of the result of step (b) to the level of resolution of said image.
A method for filtering an image having at least one high frequency feature comprising the steps of:
- (a) filtering said image with a non-linear filter targeted at said at least one high resolution feature to produce a filtered image, wherein said at least one high resolution feature is selected from a group comprising lines and points, and wherein said filtering is targeted at said at least one high resolution feature and not at edges;
- (b) resolving said filtered image into at least one lower level of resolution;
- and
- (c) analyzing said filtered image at said at least one lower level of resolution.
A method for filtering an image having at least one high frequency feature comprising the steps of:
- (a) filtering said image with a non-linear filter targeted at said at least one high resolution feature to produce a filtered image, wherein said at least one high resolution feature is selected from a group comprising lines and points, and wherein said filtering is targeted at said at least one high resolution feature and not at edges;
- (b) producing an intermediate image by subjecting said filtered image to low pass filtering;
- and
- (c) analyzing the difference between the result of step (a) and the result of step (b).
An apparatus for encoding an image having at least one high frequency feature comprising:
- (a) means for filtering said image with a non-linear filter targeted at said at least one high resolution feature to produce a filtered image, wherein said at least one high resolution feature is selected from a group comprising lines and points, and wherein said filtering is targeted at said at least one high resolution feature and not at edges;
- (b) means for low pass filtering and subsampling said filtered image into at least one lower level of resolution;
- (c) means for encoding a correction image at said at least one lower level of resolution;
- (d) means for producing an expanded result image from said correction image, said expanded result image having the same level of resolution as said image;
- and
- (e) means for encoding at the level of resolution of said image the difference between said image and said expanded result image.
Dependencies are inferred from the claim text.