Document unbending and recoloring systems and methods

Patent metadata

Field	Value
Publication number	US-9495587-B2
Application number	US-201313924010-A
Country	US
Kind code	B2
Filing date	Jun 21, 2013
Priority date	Jun 22, 2012
Publication date	Nov 15, 2016
Grant date	Nov 15, 2016

Abstract

According to one aspect, a system for processing a document image is disclosed. In an exemplary embodiment, the system includes an edge-detection unit configured to identify an edge of a document from a document image. The system also includes a keystone-correction unit and a flattening unit. The keystone-correction unit is configured to correct keystone distortion in the document image. The flattening unit is configured to flatten content of the document in the document image.

First claim

What is claimed is:

1. A system for processing a document image, comprising one or more processors, and a memory coupled to the one or more processors and storing instructions that, when executed by the one or more processors, cause the system to:
identify one or more edges of a document from the document image, wherein identifying the one or more edges of the document comprises at least one of applying an edge-finding filter to the document image and selecting a ranked potential edge of the document image, the potential edge being ranked based on one or more of a straightness of the potential edge and an intersection of the potential edge with at least one other potential edge;
correct keystone distortion in the document image; and
flatten content of the document in the document image.

2. The system of claim 1 , wherein the edge of the document image corresponds to an edge of content of the document in the document image, the content comprising at least one of text and objects.

3. The system of claim 1 , wherein applying the edge-finding filter to the document image comprises applying a convolution filter to a single channel of pixel color values of the document image.

4. The system of claim 1 , wherein a potential edge is ranked based on a combination of:
a length of the potential edge;
the straightness of the potential edge;
a vertical or horizontal orientation of the potential edge; and
the intersection of the potential edge with at least one other potential edge.

5. The system of claim 1 , wherein correcting the keystone distortion comprises:
determining a field of view of an image capture device used to capture the document image;
based on the field of view, determining at least one rotation to apply to the document image to render the document flat in the document image; and
applying the at least one rotation to the document.

6. The system of claim 1 , wherein correcting the keystone distortion comprises:
identifying a corner of the document based on edges of the document identified by the edge-detection unit, the corner formed with a first edge and second edge; and
distorting the document image based on the identified corner to render the first edge of the corner substantially perpendicular to the second edge of the corner.

7. The system of claim 6 , the memory further storing instructions that, when executed by the one or more processors, cause the system to:
after distorting the document image, determine whether the first edge or second edge in the document image corresponds to a side edge of the document; and
responsive to determining that the corresponding side edge in the document image has a curved portion, replace at least the curved portion in the document image with a substantially vertical line.

8. The system of claim 6 , the memory further storing instructions that, when executed by the one or more processors, cause the system to:
after distorting the document image, determine whether the first edge or second edge in the document image corresponds to a top edge or bottom edge of the document; and
responsive to determining that the corresponding first edge or second edge in the document image includes a curved portion, replace at least the curved portion in the document image with a substantially horizontal line.

9. The system of claim 1 , the memory further storing instructions that, when executed by the one or more processors, cause the system to:
detect that, proximate an intersection between a first identified edge and a second identified edge, the first identified edge becomes nonlinear; and
linearly extend the first identified edge to the intersection with the second identified edge.

10. The system of claim 1 , wherein the flattening comprises:
determining a virtual first horizontal line passing through a topmost vertical coordinate at a top edge of the document in the document image;
determining a virtual second horizontal line passing through a bottommost vertical coordinate at a bottom edge of the document in the document image; and
determining a linear interpolation between a downward shift from the virtual first horizontal line to the top edge of the document in the document image and an upward shift from the virtual second horizontal line to the identified bottom edge of the document in the document image.

11. The system of claim 1 , the memory further storing instructions that, when executed by the one or more processors, cause the system to: determine, via a recoloring unit, a background level for the document image based on a background level of a group of pixels in the document image.

12. The system of claim 11 , wherein determining the background level for the document image comprises determining, for the group of pixels, if comparison values of individual pixels of a subgroup of pixels within the group of pixels have a standard deviation less than predetermined threshold value.

13. The system of claim 12 , wherein the comparison values correspond to at least one of a sum and multiple of color component values for the pixels.

14. The system of claim 11 , wherein determining the background level for the document image comprises determining, from the group of pixels:
a standard deviation for each color channel across each of a plurality of subgroups of pixels; and
a total standard deviation formed from the color channel standard deviations.

15. The system of claim 11 , wherein determining the background level for the document image comprises determining a deviation of a color value associated with the group of pixels from an average color level associated with the document image.

16. A system for processing a document image, comprising one or more processors, and a memory coupled to the one or more processors and storing instructions that, when executed by the one or more processors, cause the system to:
identify one or more edges of a document from a document image;
correct keystone distortion in the document image; and
flatten content of the document in the document image, wherein the flattening comprises:
determining a virtual first horizontal line passing through a vertical coordinate at a top edge of the document in the document image;
determining a virtual second horizontal line passing through a vertical coordinate at a bottom edge of the document in the document image; and
determining a linear interpolation between a downward shift from the virtual first horizontal line to the top edge of the document in the document image and an upward shift from the virtual second horizontal line to the identified bottom edge of the document in the document image.

17. The system of claim 16 , wherein the edge of the document image corresponds to an edge of content of the document in the document image.

18. The system of claim 16 , wherein identifying the edge of the document from the document image comprises applying a filter matrix to pixel color values in the document image.

19. The system of claim 16 , wherein identifying an edge of the document from the document image comprises selecting a potential edge of the document image based on a ranking of the potential edge in comparison to other potential edges.

20. The system of claim 19 , wherein the ranking of a potential edge is based on at least one of:
a length of the potential edge;
a smoothness of the potential edge;
a vertical or horizontal orientation of the potential edge; and
an intersection of the potential edge with at least one other potential edge.

21. The sys

Assignees

• Wilson Scott E
• Hildebrandt Peter W
• Steelcase Inc

Inventors

• Wilson Scott E
• Hildebrandt Peter W

Classifications

• G06V30/414

  Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text · CPC title

• G06V10/247

  by affine transforms, e.g. correction due to perspective effects; Quadrilaterals, e.g. trapezoids · CPC title

• G06V30/40

  Document-oriented image-based pattern recognition · CPC title

• G06T2207/30176

  Document · CPC title

• G06K2009/363

  Physics · mapped topic