Iterative smoothing technique for pre-processing mixed raster content planes to improve the quality of a decompressed image and increase document compression ratios
AssigneeXerox Corporation
InventorsRicardo L. de Queiroz, Reiner Eschbach, William A. Fuss, Robert R. Buckley
Life of this patent
Estimating the term…
Abstract
A method and apparatus for compressing a mixed raster content image that represents a color or gray scale a document is disclosed. The pixel map is decomposed into a three-plane representation—a reduced-resolution “upper” plane, a reduced-resolution “lower” plane, and a high-resolution binary selector plane. An iterative smoothing technique is then used to pre-process the upper and lower planes using the information contained in the selector plane, thereby reducing the amount of data that will be subjected to further processing.
Claims12
A method of processing mixed raster content planes which represent a compressed pixel map representation of a document, wherein the mixed raster content planes include an upper plane, a lower plane, and a selector plane, and wherein the upper and lower planes have been processed based upon data contained in said selector plane, comprising the steps of:
- a) generating an image plane map which identifies locations on an associated image data plane of image data signals that shall be used to reconstruct said digital image;
- b) replacing all image data signals that correspond to locations not identified for image reconstruction using an iterative smoothing technique;
- and
- c) outputting said image plane;
- wherein for said image data signal replacing steps, said iterative smoothing technique comprises the steps of:
- 1) replacing each image data signal on said image plane that does not correspond to said reconstruction identified locations with a signal equal to an average of all image data neighbor signals that correspond to locations previously identified for reconstruction, wherein a neighbor signal is defined as an image data signal that is horizontally, vertically or diagonally adjacent to said image data signal;
- 2) applying a discrete cosine transformation to said image plane and quantizing the results of said transformation;
- 3) inverse quantizing said transformation results and performing an inverse discrete cosine transform thereon, thereby generating a pseudo-plane that has image data signals that lie in the same locations as signals in said image data plane;
- and
- 4) replacing each image data signal on said image data plane that does not correspond to a reconstruction identified location on said image data plane map with an image data signal on said pseudo-plane that lies in a same location.
A method of processing mixed master content planes which represent a compressed pixel map representation of a document, wherein the mixed raster content planes include an upper plane a lower plane and a selector plane, and wherein the upper and lower planes have been processed based upon data contained in said selector plane, comprising the steps of:
- a) inputting a threshold signal, which indicates an acceptable level of distortion for a subsequent processing operation;
- b) inputting a variance signal, which indicates a maximum acceptable magnitude difference between an image data signal and a threshold signal;
- c) generating an image plane map which identifies locations on an associated image data plane of image data signals that shall be used to reconstruct said digital image;
- d) replacing image data signals that correspond to locations not identified for image reconstruction using an iterative smoothing technique if said variance signal is greater than or equal to said threshold signal;
- and
- e) outputting said image plane;
- wherein for said image data signal replacing step, said iterative smoothing technique further comprises the steps of:
- 1) replacing each image data signal on said image plane that does not correspond to said reconstruction identified locations with a signal equal to an average of all image data neighbor signals that correspond to locations previously identified for reconstruction, wherein a neighbor signal is defined as an image data signal that is horizontally, vertically, or diagonally adjacent to said image data signal;
- 2) applying a discrete cosine transformation to said image plane and quantizing the results of said transformation;
- 3) inverse quantizing said transformation results and performing an inverse discrete cosine transform thereon, thereby generating a pseudo-plane that has image data signals that lie in the same locations as signals in said image data plane;
- and
- 4) replacing each image data signal on said image data plane that does not correspond to a reconstruction identified location on said image data plane map with an image data signal on said pseudoplane that lies in a same location.
Dependencies are inferred from the claim text.