Model-based dewarping method and apparatus

• US 8,406,476 B2
• Filed: 03/02/2007
• Issued: 03/26/2013
• Est. Priority Date: 03/02/2006
• Status: Active Grant

• Current Assignee: Compulink Management Center Incorporated
• Original Assignee: Compulink Management Center Incorporated
• Inventors: Wu, Minghui, Li, Rongfeng, Li, Wenxin, Heaney, Edward P. Jr., Chan, Karl M., Rapelje, Kurt A., Fu, Bin, Feng, Bin
• Primary Examiner(s): BHATNAGAR, ANAND P
• Application Number:
  US12/281,315
• Publication Number:
  US 20100239165A1
• Time in Patent Office: 2,216 Days
• Field of Search: 382/100, 382/103, 382/101, 382/112, 382/181, 382/182, 382/190, 382/195, 382/199, 382/206, 382/276, 382/281, 382289-292, 382/286, 382/287, 356/138, 702/150, 702/151
• US Class Current: 382/112
• CPC Class Codes: G06T 2207/30176 Document
• CPC Class Codes: G06T 3/06 Topological mapping of higher dimensional structures onto lower dimensional surfaces
• CPC Class Codes: G06T 5/00 Image enhancement or restoration
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/80 Geometric correction
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields
• CPC Class Codes: G06V 30/1478 of characters or characters lines
• CPC Class Codes: G06V 30/414 Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text
• CPC Class Codes: H04N 1/387 Composing, repositioning or otherwise geometrically modifying originals

Abstract

An apparatus and method for processing a captured image and, more particularly, for processing a captured image comprising a document. In one embodiment, an apparatus comprising a camera to capture documents is described. In another embodiment, a method for processing a captured image that includes a document comprises the steps of distinguishing an imaged document from its background, adjusting the captured image to reduce distortions created from use of a camera and properly orienting the document is described.

Claims

1. A method for processing a captured image that comprises an imaged document;

said method comprising;

separating text areas from the remainder of said captured image;

enhancing the direction of text lines in said text areas;

extracting said text lines;

selecting from said extracted text lines at least two representative text lines;

selecting a polynomial function to fit each of said representative text lines; and

determining a vanishing point based on information derived from said representive text lines.

2. The method recited in claim 1, wherein said representative text lines are the two longest lines among the extracted text lines.

3. The method recited in claim 1, wherein said representative text lines are the two most separated lines among the extracted text lines.

4. The method recited in claim 1, wherein end points of said representative text lines determine a left text edge line and a right text edge line for the imaged document.

5. The method recited in claim 4, wherein the left text edge line and the right text edge line are used to determine the vanishing point.

6. The method recited in claim 5, further comprising the step of:

forming a curved coordinate net based on the polynomial function of the representative text lines and the determined vanishing point.

7. A method for processing a captured image that comprises an imaged document;

said method comprising;

separating text areas from the remainder of said captured image;

enhancing the direction of text lines in said text areas;

extracting said text lines; and

selecting a polynomial function to fit said text lines,

wherein said enhancing step further comprising the steps of

detecting distance between said text lines;

compressing said captured image under a compression rate;

computing direction for each pixel in said text lines;

eliminating gaps between characters so that characters in the same text line connect together, but characters not in the same text line are separated.

8. The method recited in claim 7, wherein the direction for each pixel in said text of lines is determined using a plurality of 9×

9 directional templates, where each template comprises a black block in the center representing the pixel and eight gray blocks that extend from the black block that describe a direction of the pixel, and the gray level values are summed for each of the plurality of directional templates to determine the direction of the pixel.

9. The method recited in claim 8, wherein the number of the plurality of 9×

9 directional templates is 16.

10. The method recited in claim 8, after the step of computing direction for each pixel in said text lines further comprising the step of:

setting the direction of every pixels in the same vertical line to have the same direction by using a smoothing algorithm.

11. The method recited in claim 10, wherein the step of setting the direction of every pixels in the same vertical line to have the same direction is performed by a voting process.