Processing a digital image of content to remove border artifacts

• US 8,731,297 B1
• Filed: 09/28/2007
• Issued: 05/20/2014
• Est. Priority Date: 09/28/2007
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Zhou, Hanning, Liang, Jian, Yacoub, Sherif M.
• Primary Examiner(s): BEKELE, MEKONEN T
• Application Number:
  US11/864,180
• Time in Patent Office: 2,426 Days
• Field of Search: 382/103, 382/154, 382/199, 382/316, 382/276, 382/175, 382/190, 382/282, 382/254, 382/173, 382/180, 382/290, 382/275, 375/474, 375/473
• US Class Current: 382/180
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06V 10/20 Image preprocessing

Abstract

Systems and methods for removing border objects from a page of digital image are presented. More particularly, in removing border objects from a page of a digital image, the following are performed. In the first place, a digital image having at least one page of content is obtained. A content area of content on the page is determined. Border objects in the region between the page margin and the content area are then located. For each located border object, the following two steps are performed. The located border object is evaluated according to a set of border criteria, and the located border object is removed if the evaluation of the border object indicates that the border object is to be removed. Thereafter, the updated digital image is stored in a content store.

Claims

1. A method, comprising:

obtaining, in a computing device, a digital image having a page of content;

determining, in the computing device, a content area of the page of content;

locating, in the computing device, a border object in a region between a page margin and the content area;

determining whether a bounding area of the border object touches or intersects with the content area;

responsive to the bounding area of the border object touching or intersecting with the content area, determining a number of pixels of the border object that touch or intersect with the content area;

removing the border object responsive to the number of pixels of the border object that touch or intersect with the content area falling below a predetermined threshold; and

storing an updated digital image in a content store.

2. The method of claim 1, wherein the removing further comprises:

determining whether the border object is located closer to the page margin than to the content area; and

wherein the border object is removed responsive to the border object being closer to the page margin than to the content area.

3. The method of claim 1, further comprising removing the border object responsive to the border object being oblique to a nearest page margin.

4. The method of claim 1, wherein the predetermined threshold is a first predetermined threshold, and further comprising removing the border object in response to determining that a ratio of a width and a height of the border object meets a second predetermined threshold.

5. The method of claim 2, further comprising determining at least two of the following and conditionally removing the border object according to the at least two determinations:

a ratio of a width and a height of the border object, wherein the border object is removed responsive to the ratio meeting a second predetermined threshold;

whether the border object is aligned with or oblique to a nearest page margin, wherein the border object is removed responsive to being oblique to the nearest page margin; and

whether the border object is located closer to the page margin than the content area, wherein the border object is removed responsive to being closer to the page margin.

6. The method of claim 1, further comprising:

determining whether the border object touches a content in the content area;

responsive to the border object touching content in the content area, determining a number of pixels of the border object that touch the content in the content area; and

removing the border object responsive to the number of pixels of the border object that touch the content in the content area falling below a second predetermined threshold.

7. The method of claim 1, wherein locating the border object in the region between the page margin and the content area comprises:

identifying a pixel in the region between the page margin and the content area that has not been identified as belonging to another border object; and

applying a connected component algorithm beginning with the pixel to determine an extent of the border object in which the pixel is included.

22. The method of claim 1, wherein removing the border object comprises:

segmenting the digital image into a plurality of layers, wherein individual ones of the plurality of layers correspond to at least one of the image, a text region or a text line; and

despeckling the individual ones of the layers using a plurality of median filters, wherein a respective one of the median filters used is based at least in part on a content of a respective one of the layers.

27. The method of claim 1, further comprising de-skewing, in the computing device, the digital image prior to locating the border object.

28. The method of claim 1, further comprising removing the border object responsive to a proximity of the border object with respect to the page margin indicating that the border object should be removed.

8. A computer-implemented method, comprising:

obtaining, in a computing device, a page of a digital image;

identifying, in the computing device, a border object that falls within a border area of the page, the border area comprising an area of the page that lies between a page margin and a content area that bounds a content of the page;

determining whether a bounding area for the border object touches or intersects with the content area;

determining, responsive to the bounding area for the border object touching or intersecting with the content area, a number of pixels of the border object that touch or intersect with the content area;

retaining the border object responsive to the number of pixels of the border object that touch or intersect with the content area meeting a predetermined threshold; and

storing the digital image in a content store.

9. The method of claim 8, wherein the predetermined threshold is a first predetermined threshold, and further comprising retaining the border object responsive to at least one of:

a ratio of a width and a height of the border object falling below a second predetermined threshold;

the border object being aligned with a nearest page margin; and

the border object being closer to the content area than the page margin.

10. The method of claim 8, wherein the predetermined threshold is a first predetermined threshold, and further comprising determining whether the border object touches the content of the page in the content area;

responsive to the border object touching the content of the page in the content area, determining a number of pixels of the border object that touch the content of the page in the content area; and

retaining the border object responsive to the number of pixels of the border object that touch the content of the page in the content area meeting a second predetermined threshold.

11. The method of claim 8, wherein identifying the border object that falls within the border area of the page comprises:

identifying a pixel in the border area of the page that has not been identified as belonging to another border object; and

applying a connected component algorithm, beginning with the pixel, to determine an extent of the border object in which the pixel is included.

23. The computer-implemented method of claim 8, wherein deleting the border object comprises:

segmenting the digital image into a plurality of layers, wherein individual ones of the plurality of layers correspond to at least one of the image, a text region or a text line; and

despeckling individual ones of the layers using a plurality of median filters, wherein a respective one of the median filters used is based on a content of a respective one of the layers.

12. A system, comprising:

a processor;

a memory;

a content store capable of storing a digital image of printed content, the digital image having at least one page of content; and

wherein the system is executable using the processor and the memory, and is configured to obtain the digital image to process, and for individual ones of the at least one page of content in the digital image;

identify a border object that falls within a border area of the at least one page, the border area comprising an area of the at least one page that lies between a page margin and a content area that bounds the content;

determine whether a bounding area for the border object touches or intersects with the content area;

determine, responsive to the bounding area touching or intersecting with the content area, a number of pixels of the border object that touch or intersect with the content area;

retain the border object responsive to the number of pixels of the border object that touch or intersect with the content area meeting a predetermined threshold; and

store the digital image in the content store.

13. The system of claim 12, wherein the predetermined threshold is a first predetermined threshold, and the system is configured to retain the border object responsive to:

a ratio of a width and a height of the border object falling below a second predetermined threshold;

the border object being aligned with a nearest page margin; and

the border object being located closer to the page margin than the content area.

14. The system of claim 12, wherein the predetermined threshold is a first predetermined threshold, and the system is further configured to:

determine whether the border object touches the content;

responsive to the border object touching the content, determine a number of pixels of the border object that touch the content; and

retain the border object responsive to the number of pixels of the border object that touch the content meeting a second predetermined threshold.

15. The system of claim 12, wherein identifying the at least one border object that falls within the border area of the page further comprises:

identifying a pixel in the border area of the page that has not been identified as belonging to another border object; and

applying a connected component algorithm, beginning with the pixel, to determine an extent of a border object in which the pixel is included.

24. The system of claim 12, wherein deleting the border object comprises:

segmenting the digital image into a plurality of layers, wherein individual ones of the plurality of layers correspond to at least one of the image, a text region or a text line; and

despeckling the individual ones of the layers using a plurality of median filters, wherein a respective one of the median filters used is based on a content of a respective one of the layers.

29. The system of claim 12, wherein the system is further configured to:

evaluate the border object according to a border criteria; and

retain the border object in the page responsive to the evaluating indicating that the border object should be retained.

16. A system, comprising:

a processor;

a memory; and

a content store capable of storing a digital image of printed content, the digital image having at least one page of content;

wherein the system is executable using the processor and the memory, and is configured to;

obtain the digital image to process

for individual ones of the at least one page of content in the digital image;

determine a content area for a respective page of content;

locate a border object in a border region between a page margin and the content area;

determine whether a bounding area for the border object touches or intersects with the content area;

determine, responsive to the bounding area touching or intersecting with the content area, a number of pixels of the border object that touch or intersect with the content area;

remove the border object responsive to the number of pixels of the border object that touch or intersect with the content area falling below a predetermined threshold; and

store the digital image in the content store.

17. The system of claim 16, wherein locating the border object in the border region between the page margin and the content area further comprises:

identifying a pixel in the border region of the page that has not been identified as belonging to another border object; and

applying a connected component algorithm beginning with the pixel, to determine an extent of the border object in which the pixel is included.

18. The system of claim 16, wherein the predetermined threshold is a first predetermined threshold, and the system is further configured to remove the border object responsive to at least one of:

a ratio of a width and a height of the border object meeting a second predetermined threshold; and

the border object being oblique to a nearest page margin.

25. The system of claim 16, wherein removing the border object comprises:

segmenting the digital image into a plurality of layers, wherein individual ones of the plurality of layers correspond to at least one of the image, a text region or a text line; and

despeckling the individual ones of the layers using a plurality of median filters, wherein a respective one of the median filters used is based on a content of a respective one of the layers.

30. The system of claim 16, wherein the system is further configured to remove the border object responsive to the border object being closer to the page margin than the content area.

19. A non-transitory computer-readable medium bearing computer-executable instructions which, when executed on a computing device having a processor and a memory, carry out a method comprising:

obtaining a page of a digital image;

identifying a border object that falls within a border area of the page, the border area comprising an area of the page that lies between a page margin and a content area that bounds a content of the page;

determining whether a bounding area for the border object touches or intersects with the content area;

determining, responsive to the bounding area touching or intersecting with the content area, a number of pixels of the border object that touch or intersect with the content area;

deleting the border object responsive to the number of pixels of the border object that touch or intersect with the content area falling below a threshold; and

storing the digital image in a content store.

20. The non-transitory computer-readable medium of claim 19, wherein identifying the border object that falls within the border area of the page comprises:

identifying a pixel in the border area of the page that has not been identified as belonging to another border object; and

applying a connected component algorithm, beginning with the pixel, to determine an extent of a border object in which the pixel is included.

21. The non-transitory computer-readable medium of claim 19, wherein the threshold is a first threshold, and the method further comprises determining:

whether a ratio of a width and a height of the border object meets a second threshold, wherein the border object is deleted responsive to the ratio meeting the second threshold;

whether the border object is aligned with or oblique to a nearest page margin, wherein the border object is deleted responsive to being oblique to the nearest page margin; and

whether the border object is located closer to the page margin than the content area, wherein the border object is deleted responsive to being closer to the page margin than the content area.

26. The non-transitory computer-readable medium of claim 19, wherein deleting the border object comprises:

segmenting the digital image into a plurality of layers, wherein individual ones of the plurality of layers correspond to at least one of the image, a text region or a text line; and

despeckling the individual ones of the layers using a plurality of median filters, wherein a respective one of the median filters used is based on a content of a respective one of the layers.