Method for detecting a document boundary

• US 9,122,921 B2
• Filed: 04/28/2014
• Issued: 09/01/2015
• Est. Priority Date: 06/12/2013
• Status: Active Grant

• Current Assignee: Kodak Alaris Inc (KPP (no. 2) Trustees Ltd.)
• Original Assignee: Kodak Alaris Inc (KPP (no. 2) Trustees Ltd.)
• Inventors: Beato, Louis James, Alescio, John J.
• Primary Examiner(s): Safaipour, Houshang
• Application Number:
  US14/263,181
• Publication Number:
  US 20140368891A1
• Time in Patent Office: 491 Days
• Field of Search: 358/518, 358/505, 358/474, 358/498
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10008 from scanner, fax or copier
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30176 Document
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 10/243 by compensating for image skew or non-uniform image deformations
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition
• 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 1/00748 Detecting edges, e.g. of a stationary sheet

Abstract

A method for detecting a document boundary in a captured digital image depicting a hardcopy document on a background. Each color channel of the captured digital image is analyzed to determine a corresponding busyness metric representing a complexity level of the image data. The color channel having a lowest busyness level is selected and analyzed to detect a document boundary of the depicted hardcopy document. The detected document boundary can be used to perform a perspective correction process to determine a corrected digital image where the depicted document has a substantially rectangular boundary.

Claims

1. A method for detecting a document boundary in a captured digital image of a document, the method implemented at least in part by a data processing system and comprising:

receiving a color digital image captured using a digital imaging system that depicts a hardcopy document on a background, wherein the color digital image includes image data for a plurality of color channels;

for each of the color channels analyzing the image data to determine a corresponding busyness metric, the busyness metric representing a complexity level of the image data, wherein the determination of the busyness metric includes determining a compressibility metric for the image data;

selecting the color channel having a lowest busyness level;

analyzing the image data for the selected color channel to detect the document boundary of the depicted hardcopy document; and

storing an indication of the detected document boundary in a processor-accessible memory.

2. The method of claim 1 further including thresholding the image data for each of the color channels prior to computing the busyness metric.

3. The method of claim 2 wherein the thresholding of the image data includes:

analyzing the image data to determine an adaptive threshold; and

using the adaptive threshold to threshold the image data.

4. The method of claim 3 wherein the adaptive threshold is determined based on statistical properties of the image data.

5. The method of claim 2 wherein the determination of the busyness metric includes counting a number of transitions along a set of paths through the thresholded image data.

6. The method of claim 2 wherein the determination of the busyness metric includes performing a connected component analysis of the thresholded image data.

8. The method of claim 1 wherein the determination of the busyness metric includes determining an image entropy metric for the image data.

9. The method of claim 1 wherein the determination of the busyness metric includes determining a frequency content metric for the image data.

10. The method of claim 1 wherein detection of the document boundary includes performing an edge detection process to detect edges in the image data.

11. The method of claim 10 further including determining one or more polygons corresponding to the detected edges.

12. The method of claim 11 wherein a plurality of polygons are determined corresponding to the detected edges and a single polygon is selected based on one or more predefined selection criteria.

13. The method of claim 11 wherein a selection criterion discards polygons that are not quadrilaterals.

14. The method of claim 11 wherein a selection criterion discards polygons that do not have a minimum size.

15. The method of claim 1 wherein the color channels include a color channel that represents a hue attribute, a color channel that represents a saturation attribute and a color channel that represents a luminance attribute.

16. The method of claim 1 wherein the color channels include a color channel that represents a luminance attribute and two color channels representing color difference attributes.

17. The method of claim 1 wherein the color channels include red, green and blue color channels.

7. A method for detecting a document boundary in a captured digital image of a document, the method implemented at least in part by a data processing system and comprising:

receiving a color digital image captured using a digital imaging system that depicts a hardcopy document on a background, wherein the color digital image includes image data for a plurality of color channels;

for each of the color channels analyzing the image data to determine a corresponding busyness metric, the busyness metric representing a complexity level of the image data;

selecting the color channel having a lowest busyness level;

analyzing the image data for the selected color channel to detect the document boundary of the depicted hardcopy document; and

storing an indication of the detected document boundary in a processor-accessible memory; and

applying a perspective correction process to the captured color digital image based on the detected document boundary to determine a corrected digital image where the depicted document has a substantially rectangular boundary.

20. The method of claim 7 wherein the determination of the busyness

metric includes determining a compressibility metric for the image data.

18. A document imaging system for capturing a digital image of a document, comprising:

an image capture system adapted to capture a color digital image of a hardcopy document on a background, wherein the captured color digital image includes image data for a plurality of color channels;

a data processing system; and

a memory system communicatively connected to the data processing system and storing instructions configured to cause the data processing system to implement a method for correcting perspective distortion in the captured color digital image, wherein the method includes;

transforming the captured color digital image to an analysis color space having a plurality of analysis color channels;

for each of the analysis color channels analyzing the image data to determine a corresponding busyness metric, the busyness metric representing a complexity level of the image data;

selecting the analysis color channel having a lowest busyness level;

analyzing the image data for the selected analysis color channel to detect the document boundary of the depicted document;

applying a perspective correction process to the captured color digital image based on the detected document boundary to determine a corrected digital image where the depicted document has a substantially rectangular boundary; and

storing the corrected digital image in a processor-accessible memory.

19. A method for detecting a document boundary in a captured digital image of a document, the method implemented at least in part by a data processing system and comprising:

receiving a color digital image captured using a digital imaging system that depicts a hardcopy document on a background, wherein the color digital image includes image data for a plurality of color channels;

for each of the color channels analyzing the image data to determine a corresponding busyness metric, the busyness metric representing a complexity level of the image data, wherein the determination of the busyness metric includes determining an image entropy metric for the image data;

selecting the color channel having a lowest busyness level;

analyzing the image data for the selected color channel to detect the document boundary of the depicted hardcopy document; and

storing an indication of the detected document boundary in a processor-accessible memory.