Method and apparatus for detection of a skew angle of a document image using a regression coefficient

• US 5,563,403 A
• Filed: 12/14/1994
• Issued: 10/08/1996
• Est. Priority Date: 12/27/1993
• Status: Expired due to Term

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Bessho, Goroh, Ejiri, Koichi
• Primary Examiner(s): Epps, Georgia Y.
• Assistant Examiner(s): Steady, Jacqueline M.
• Application Number:
  US08/355,902
• Time in Patent Office: 664 Days
• Field of Search: 250/208.1, 250/559.36, 250/559.37, 358/474, 358/488, 358/494, 382/289, 382/290, 382/278, 382/168, 382/171
• US Class Current: 250/208.1
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1478 of characters or characters lines

Abstract

A method and apparatus for detecting the skew angle of a document image. Skew angle determination is performed by the steps of determining a set of sampling points from an input document image and processing X and Y coordinates of the sampling points in order to calculate a regression coefficient of the sampling points. The skew angle of the document is determined using the regression coefficient. To evaluate a calculated skew angle which corresponds to the regression coefficient, a correlation coefficient is calculated and evaluated. As coordinates of sampling points are obtained for a plurality of sets of data corresponding to different ruled lines or lines of characters, a histogram may be used to determine the most probable skew angle.

Claims

1. A method for detecting a skew angle, comprising the steps of:

scanning a page of a document and generating a document image;

determining sampling points of the document image using ruled lines of the document image;

calculating a regression coefficient of the sampling points; and

determining the skew angle of the document image using said regression coefficient.

2. A method as recited in claim 1, further comprising the steps of:

calculating a correlation coefficient using the sampling points; and

evaluating the regression coefficient using the correlation coefficient of said sampling points.

3. A method as recited in claim 1, wherein said step of determining a skew angle comprises:

calculating a histogram using a plurality of calculated regression coefficients; and

determining the skew angle using the calculated histogram.

4. A method as recited in claim 3, further comprising the step of:

calculating a correlation coefficient of a plurality of sets of sampling data using the sampling points;

wherein the step of calculating the histogram calculates a weighted histogram using the correlation coefficients of the plurality of sets of sampling points.

5. A method as recited in claim 1, wherein the scanning step includes:

irradiating the page of the document with a light source.

6. A method as recited in claim 1, wherein the step of determining sampling points determines sampling points which are located at a boundary of a black pixel and a white pixel of the document image.

7. A method as recited in claim 1, wherein said step of determining sampling points determines the sampling points using characters of the document image.

8. A method for detecting a skew angle, comprising the steps of:

scanning a page of a document and generating a document image;

determining if at least a predetermined number of adjacent pixels forming a ruled line exist;

when at least the predetermined number of adjacent pixels forming a line are determined not to exist, performing the steps of;

determining sampling points using characters of the document image;

calculating a regression coefficient using the determined sampling points; and

calculating a skew angle using the regression coefficient.

9. A method according to claim 8, wherein when at least the predetermined number of adjacent pixels forming a line are determined to exist, performing the steps of:

determining sampling points of a ruled line of the document image which correspond to the adjacent pixels;

calculating a regression coefficient using the determined sampling points of the ruled line; and

calculating a skew angle using the regression coefficient calculated using the sampling points of the ruled line.

10. A method according to claim 9, further comprising, which are performed after the step of determining the sampling points of the ruled line, the steps of:

calculating a correlation coefficient using the determined sampling points of the ruled line;

determining if the correlation coefficient is less than a predetermined threshold;

performing the steps which are performed when at least the predetermined number of adjacent pixels forming a line are determined not to exist, when the correlation coefficient is determined to be less than the predetermined threshold.

11. A method according to claim 10, wherein the steps of calculating the regression coefficient using the determined sampling points of the ruled line and the calculating the skew angle using the regression coefficient calculated using the sampling points of the ruled line are performed only when the correlation coefficient is determined to be greater than or equal to the predetermined threshold.

12. A system for detecting a skew angle, comprising:

means for determining sampling points of a document image using ruled lines of the document image;

means for calculating a regression coefficient of the sampling points; and

means for determine the skew angle of the document image using said regression coefficient.

13. A system as recited in claim 12, further comprising:

means for calculating a correlation coefficient using the sampling points; and

means for evaluating the regression coefficient using the correlation coefficient of said sampling points.

14. A system as recited in claim 12, wherein said means for determining a skew angle comprises:

means for calculating a histogram using a plurality of calculated regression coefficients; and

means for determining the skew angle using the calculated histogram.

15. A system as recited in claim 14, further comprising:

means for calculating a correlation coefficient of a plurality of sets of sampling data using the sampling points;

wherein the means for calculating the histogram calculates a weighted histogram using the correlation coefficients of the plurality of sets of sampling data.

16. A system as recited in claim 12, further comprising a scanner for generating the document image;

irradiating the page of the document with a light source.

17. A system as recited in claim 12, wherein the means for determining sampling points determines sampling points which are located at a boundary of a black pixel and a white pixel of the document image.

18. A system as recited in claim 12, wherein said means for determining sampling points determines the sampling points using characters of the document image.

19. A method for detecting a skew angle, comprising the steps of:

scanning a page of a document and generating a document image;

determining if a skew angle of the document image can be determined using ruled lines of the document image;

determining the skew angle of the document image using the ruled lines of the document image, when the skew angle of the document image can be determined using ruled lines of the document image; and

determining the skew angle of the document image using character information, when the skew angle of the document image cannot be determined using the ruled lines of the document image.

20. A system for detecting a skew angle, comprising:

a scanning device which scans a page of a document and generates a document image;

means for determining if a skew angle of the document image can be determined using ruled lines of the document image;

means for determining the skew angle of the document image using the ruled lines of the document image, when the skew angle of the document image can be determined using ruled lines of the document image; and

means for determining the skew angle of the document image using character information, when the skew angle of the document image cannot be determined using the ruled lines of the document image.

21. A method for detecting a skew angle, comprising the steps of:

scanning a page of a document and generating a document image;

determining coordinates for characters of the document image which correspond on a one-to-one basis with the characters of the page;

determining sampling points of the document image using the coordinates which have been determined;

calculating a regression coefficient of the sampling points; and

determining the skew angle of the document image using said regression coefficient.

22. A method according to claim 21, wherein said step of determining coordinates comprises:

determining rectangles, which correspond on a one-to-one basis with the characters, the rectangles circumscribing each of the characters; and

determining the coordinates of the characters using the rectangles.

23. A method according to claim 22, wherein the step of determining the coordinates of the characters using the rectangles includes:

determining the coordinates of the characters using a corner of the rectangles.

24. A method according to claim 23, wherein the step of determining the coordinates of the characters using the rectangles includes:

determining the coordinates of the characters using an upper left-hand corner of the rectangles.

25. A system for detecting a skew angle of a page, comprising:

means for determining coordinates for characters of the page which correspond on a one-to-one basis with the characters of the page;

means for determining sampling points of the page using the coordinates which have been determined;

means for calculating a regression coefficient of the sampling points; and

means for determine the skew angle of the page using said regression coefficient.

26. A system according to claim 25, wherein the means for determining coordinates comprises:

means for determining rectangles, which correspond on a one-to-one basis with the characters, the rectangles circumscribing each of the characters; and

means for determining the coordinates of the characters using the rectangles.

27. A system according to claim 26, wherein the means for determining the coordinates of the characters using the rectangles includes:

means for determining the coordinates of the characters using a corner of the rectangles.

28. A system according to claim 27, wherein the means for determining the coordinates of the characters using the rectangles includes:

means for determining the coordinates of the characters using an upper left-hand corner of the rectangles.