Block retouching

• US 7,512,274 B2
• Filed: 03/30/2006
• Issued: 03/31/2009
• Est. Priority Date: 04/25/2002
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Malvar, Henrique S., Simard, Patrice Y.
• Primary Examiner(s): Chawan; Sheela C
• Application Number:
  US11/393,297
• Publication Number:
  US 20060171604A1
• Time in Patent Office: 1,097 Days
• Field of Search: 382/195, 382/232, 382/166, 382/269, 382/236, 382/173, 382/233, 382/164, 382/240, 382/266, 382/284, 382/154, 382/264, 345/629, 345/419, 345/634, 345/611, 345/428, 348/586, 348/587, 348/592, 348/590, 348/597, 348/E5.052, 348/699, 348/E5.059, 348/E5.058, 348/E9.056, 348/578, 358/464, 358/448, 358/450, 375/E7.081, 375/E7.04, 375/E7.166, 375/E7.112, 375/E7.182, 375/E7.088, 375/E7.19, 375/E7.226
• US Class Current: 382/195
• CPC Class Codes: G06T 5/77 Retouching; Inpainting; Scratch removal
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation
• CPC Class Codes: H04N 19/60 using transform coding
• CPC Class Codes: H04N 19/86 involving reduction of coding artifacts, e.g. of blockiness

Abstract

A system and method facilitating image retouching is provided. The invention includes an image retoucher having a boundary detector and an image extender. The invention provides for the image retoucher to extend care pixels of at least one of a foreground and a background near a detected spurious boundary by altering the binary mask used for compression of the foreground and/or the background.

Claims

1. A document compression system, comprising:

an encoder that compresses at least one of a foreground and a background of a document image;

a mask separator that processes the document image and outputs a binary mask, the binary mask assigns pixels of the document image into at least one of the foreground and the background;

an image retoucher comprising a boundary detector that detects a spurious boundary in the document image comprising the foreground and the background, detection being based, at least in part, upon pixel assignment information stored in the binary mask, the image retoucher further comprising an image extender that extends care pixels of at least one of the foreground and the background near the spurious boundary, extension of the care pixels being used for compression of at least one of the foreground and the background; and

a decoder that un-compresses at least one of the foreground and the background of the document image.

2. The system of claim 1, the boundary detector detects the spurious boundary by identifying a horizontal or vertical boundary line between merged regions that is longer than a first threshold quantity of pixels for which one side of the boundary line is foreground and the other side of the boundary is background.

3. The system of claim 2, the boundary detector further performs a linear regression along the boundary line, the spurious boundary being detected if the linear regression on either side differs by more than a second threshold amount.

4. The system of claim 3, extension of the care pixels being based upon extending care pixels near the spurious boundary of the foreground on the side of the spurious boundary that is foreground and extending care pixels near the spurious boundary of the background on the side of the spurious boundary that is background.

5. The system of claim 2, the boundary detector further performs a constant regression along the boundary line, wherein the average of the care pixels immediately against the boundary line determines the spurious boundary.

6. The system of claim 2, the boundary detector further compares the average of the foreground along the boundary line with the average of the background on the other side of the boundary, the spurious boundary being detected if the averages differ by more than a second threshold amount.

7. The system of claim 1, the binary mask being based, at least in part, upon a partition of the document image based, at least in part, upon minimization of pixel energy of at least one of the foreground and the background.

8. A method for facilitating document compression of an image, comprising:

compressing at least one of a foreground and a background of a document image;

processing the document image and outputting a binary mask;

assigning pixels of the document image into at least one of the foreground and the background;

detecting a spurious boundary in the document image comprising the foreground and the background, detection being based, at least in part, upon pixel assignment information;

extending care pixels of at least one of the foreground and the background near the spurious boundary, extension of the care pixels being used for compressing at least one of the foreground and the background; and

un-compressing at least one of the foreground and the background of the document image.

9. The method of claim 8, detection of the spurious boundary being based, at least in part, upon identification of a horizontal or vertical boundary line between merged regions that is longer than a first threshold quantity of pixels for which one side of the boundary line is foreground and the other side of the boundary is background.

10. The method of claim 9, detection of the spurious boundary being based, at least in part, upon a linear regression along the boundary line, the spurious boundary being detected if the linear regression on either side differs by more than a second threshold amount.

11. The method of claim 9, detection of the spurious boundary being based, at least in part, upon a constant regression along the boundary line, the spurious boundary being detected based on an average of the care pixels immediately against the boundary line.

12. The method of claim 9, detection of the spurious boundary being based, at least in part, upon an average of the foreground along the boundary line with an average of the background on the other side of the boundary, the spurious boundary being detected if the averages differ by more than a second threshold amount.

13. The method of claim 8, further comprising reconstructing the document image using the binary mask, the binary mask is used to determine which pixels came from the foreground and which pixels came from the background;

and altering information stored in the binary mask to reduce compression errors near the spurious boundary while keeping the binary mask the same.

14. An image retoucher, comprising:

a boundary detector that detects a spurious boundary in a document image comprising a foreground and a background, detection being based, at least in part, upon identifying a horizontal or vertical boundary line between merged regions that is longer than a first threshold quantity of pixels for which one side of the boundary line is foreground and the other side of the boundary is background, the boundary detector further performs a linear regression along the boundary line, the spurious boundary being detected if the linear regression on either side differs by more than a second threshold amount; and

an image extender that extends care pixels of at least one of the foreground and the background near the spurious boundary, extension of the care pixels being used for compression of at least one of the foreground and the background.

15. The image retoucher of claim 14, extension of the care pixels being based upon extending care pixels near the spurious boundary of the foreground on the side of the spurious boundary that is foreground and extending care pixels near the spurious boundary of the background on the side of the spurious boundary that is background.

16. The image retoucher of claim 14, the boundary detector further performs a constant regression along the boundary line, wherein the average of the care pixels immediately against the boundary line determines the spurious boundary.

17. The image retoucher of claim 14, the boundary detector further compares the average of the foreground along the boundary line with the average of the background on the other side of the boundary, the spurious boundary being detected if the averages differ by more than a second threshold amount.

18. The image retoucher of claim 14, wherein the boundary detector reconstructs the document image using a binary mask, the binary mask is used to determine which pixels came from the foreground and which pixels came from the background.

19. The image retoucher of claim 18, wherein the boundary detector alters care pixels of at least one of the foreground and the background stored in the binary mask to reduce compression errors near the spurious boundary while keeping the binary mask the same.

20. The image retoucher of claim 18, the binary mask being based, at least in part, upon a partition of the document image based, at least in part, upon minimization of pixel energy of at least one of the foreground and the background.