Detecting and utilizing add-on information from a scanned document image

• US 20020102022A1
• Filed: 01/31/2001
• Published: 08/01/2002
• Est. Priority Date: 01/31/2001
• Status: Active Grant

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Ma, Yue, Guo, Jinhong Katherine
• Granted Patent: US 6,909,805 B2
• Time in Patent Office: Days
• US Class Current: 382/170
• CPC Class Codes: G06V 30/413 Classification of content, e.g. text, photographs or tables

Abstract

A scanned document image, including add-on information such as handwritten annotations in addition to printed text lines, is processed by a handwriting detection method. First, at least one projection histogram is generated from the scanned document image. A regular pattern that correlates to the printed text lines is determined from the projection histogram. Second, connected component analysis is applied to the scanned document image to generate at least one merged text line. Each merged text line relates to at least one of the handwritten annotation and the printed text line. By comparing the merged text lines to the regular pattern of the projection histograms, the printed text lines are discriminated from the handwritten annotations.

Claims

1. A method for detecting handwritten annotations from a scanned document image, the scanned document image having a handwritten annotation and at least one printed text line, the method comprising the steps of:

generating at least one projection histogram for the scanned document image, wherein the projection histogram includes a regular pattern that correlates to the printed text lines;

applying connected component analysis to the scanned document image in order to generate at least one merged text line, wherein each of the merged text lines correlate to at least one of the handwritten annotations and the printed text line; and

comparing the merged text lines to the regular pattern of the projection histograms, thereby discriminating the printed text lines from the handwritten annotations.

2. The method of claim 1, wherein the regular pattern is margins of the printed text line.

3. The method of claim 1, wherein the projection histogram is selected from the group consisting of a vertical projection histogram, a horizontal projection histogram and the combinations thereof.

4. The method of claim 1 further comprising the step of detecting text line peaks from the projection histogram of the scanned document image.

5. The method of claim 1, wherein the step of applying connected component analysis further comprising steps of:

generating connected components onto the scanned document image by connecting dark pixels that are in relation to each other;

generating bounding boxes of the connected components; and

line merging the bounding boxes that are within a same text line in order to generate the merged text line.

6. The method of claim 5, wherein the step of line merging is recursively performed until there is no bounding box on the same text line to line merge.

7. The method of claim 5, wherein the step of line merging merges two of the bounding boxes if one bounding box is completely inside the other bounding box.

8. The method of claim 5, wherein the step of line merging merges two of the bounding boxes if two bounding boxes are neighbors and on a same line.

9. The method of claim 1 further comprising the step of separating the printed text lines and the handwritten annotations by comparing the size of the bounding box to the size of the printed text lines.

10. A method for detecting handwritten annotations from a scanned document image, the scanned document image having a handwritten annotation and at least one printed text line, the method comprising the steps of:

generating vertical and horizontal projection histograms for the scanned document image, wherein the projection histogram includes margins that correlate to the printed text lines;

generating connected components by connecting dark pixels that are in association with the others on the scanned document image;

generating bounding boxes that encapsulates all of the connected components that are in relation with each other;

line merging the bounding boxes that are within a same text line in order to generate at least one merged text line, wherein each of the merged text lines correlates to at least one of the handwritten annotations and the printed text lines;

detecting text line peaks from the horizontal projection histogram of the scanned document image; and

comparing the merged text lines to the margins determined from the vertical and horizontal projection histograms, thereby discriminating the printed text lines from the handwritten annotation.

11. A method for recording a history of a document comprising the steps of:

scanning a document having printed text lines and handwritten annotations;

separating the handwritten annotations from the printed text lines of the scanned document;

comparing the scanned document with an original document wherein the original document includes only the printed text lines;

determining an existence of previous versions of the document; and

recording a history of the document based on the separated handwritten annotations.

12. The method of claim 11 wherein the step of separating the handwritten annotations from the printed text lines of the scanned document further comprising the steps of:

generating at least one projection histogram for the scanned document image, wherein the projection histogram includes a regular pattern that correlates to the printed text lines;

applying connected component analysis to the scanned document image in order to generate at least one merged text line, wherein each merged text line relates to at least one of the handwritten annotations and the printed text line; and

comparing the merged text lines to the regular pattern of the projection histograms, thereby discriminating the printed text lines from the handwritten annotations.

13. The method of claim 12, wherein the step of applying connected component analysis further comprising steps of:

generating connected components by connecting dark pixels that are in relation to each other on the scanned document image;

generating bounding boxes of the connected components; and

line merging the bounding boxes that are within a same text line in order to generated the merged text line.

14. The method of claim 11 further comprising the steps of:

detecting the existence of the original document;

storing the printed text lines as the original document when the printed text lines of the scanned document are different from the detected original document; and

recording the scanned document as a second version of the original document.

15. The method of claim 11, wherein the step of recording the history of the document further comprising the steps of:

recording the scanned document as a second version of the original document if no previous version of the original document is detected; and

recording the scanned document as a latest version of the original document if at least one previous version is detected.

17. The method of claim 16, wherein the step of separating the printed text lines from the handwritten annotations further comprising the steps of:

generating vertical and horizontal projection histograms for the scanned document image, wherein the projection histogram includes margins that correlate to the printed text lines;

generating connected components by connecting dark pixels that are in association with the others on the scanned document image;

generating bounding boxes that encapsulates all of the connected components that are in relation with each other;

line merging the bounding boxes that are within a same text line in order to generate at least one merged text line, wherein each of the merged text lines correlates to at least one of the handwritten annotations and the printed text lines;

detecting text line peaks from the horizontal projection histogram of the scanned document image; and

comparing the merged text lines to the margins determined from the vertical and horizontal projection histograms, thereby discriminating the printed text lines from the handwritten annotations.

18. The method of claim 17, wherein the step of transmitting the document further comprising the steps of:

transmitting the original version when the printed text lines are identical to the original version of the document; and

transmitting the printed text lines when the printed text lines are different from the original version of the document.

20. The method of claim 19, wherein the step of separating the printed text lines from the handwritten annotations further comprising the steps of:

generating vertical and horizontal projection histograms for the scanned document image, wherein the projection histogram includes margins that correlate to the printed text lines;

generating connected components by connecting dark pixels that are in association with the others on the scanned document image;

generating bounding boxes that encapsulates all of the connected components that are in relation with each other;

line merging the bounding boxes that are within a same text line in order to generate at least one merged text line, wherein each of the merged text lines correlates to at least one of the handwritten annotations and the printed text lines;

detecting text line peaks from the horizontal projection histogram of the scanned document image; and

comparing the merged text lines to the margins determined from the vertical and horizontal projection histograms, thereby discriminating the printed text lines from the handwritten annotations.

21. The method of claim 19, wherein the step of storing the handwritten annotations further includes a step of linking the handwritten annotation to the original version in the database.

16. A method for securing transmission of an original version of a document, the document having the original version and at least one secondary version, wherein the original version includes only printed text lines and the secondary version includes the printed text lines and handwritten annotations, the method comprising the steps of:

separating the printed text lines from the handwritten annotations in at least one secondary version of the document;

determining the original version of the document in relation with the printed text lines separated from the handwritten annotations; and

transmitting the printed text lines after the step of determining the original version of the document.

19. An efficient compression method wherein an original version of a document is stored in a database, the document having at least one secondary version which includes printed text lines and handwritten annotations, the method comprising the steps of:

separating the printed text lines from the handwritten annotations in at least one secondary version of the document;

comparing the printed text lines with the original documents; and

storing only the handwritten annotations in association with the original document.