Systems and methods for automatic form segmentation for raster-based passive electronic documents

• US 20040194035A1
• Filed: 03/31/2003
• Published: 09/30/2004
• Est. Priority Date: 03/31/2003
• Status: Active Grant

• Current Assignee: Siemens Corp. (Siemens AG)
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Chakraborty, Amit
• Granted Patent: US 7,305,612 B2
• Time in Patent Office: Days
• US Class Current: 715/234
• CPC Class Codes: G06F 40/134 Hyperlinking
• CPC Class Codes: G06F 40/143 Markup, e.g. Standard Generalized Markup Language [SGML] or Document Type Definition [DTD]
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting

Abstract

Systems and methods for automatically extracting form information (document structure, elements, format, etc.) from electronic documents such as raster-based passive documents, and storing such form information in a file in accordance with a predetermined DTD (document type definition).

Claims

1. A method for processing electronic documents, comprising the steps of:

receiving as input an electronic document, wherein at least a portion of the electronic document is raster-based;

extracting form information from text portions and non-text portions of the electronic document; and

generating a structured document representing the extracted form information for the electronic document based on a predefined document type definition.

2. The method of claim 1, wherein the step of extracting form information comprises the steps of:

segmenting text portions and non-text portions;

separately processing the text portions and non-text portions to extract associated form information; and

combining the extracted form information of the text portions and non-text portions.

3. The method of claim 1, wherein the electronic document comprises a PDF (portable document format) file.

4. The method of claim 2, wherein the step of processing the text portions of the electronic document comprises:

extracting text segments and corresponding location information of the extracted text segments; and

determining a context of each text segment using the location information.

5. The method of claim 4, wherein the context of a text segment comprises one of a paragraph, heading, title and subheading.

6. The method of claim 4, further comprising the steps of extracting form lines and extracting table boxes.

7. The method of claim 2, wherein the step of processing the non-text portions comprises the steps of:

extracting images and location information for the images;

converting the extracted images to black and white images;

determining horizontal and vertical lines in the images; and

determining table entries and form fields using the horizontal and vertical lines.

8. The method of claim 7, wherein the step of determining horizontal and vertical lines comprises the steps of:

performing a Hough Transformation to extract line information from an image;

determining lines in the image using the extracted line information; and

discarding lines having a slope that exceeds a predetermined threshold.

9. The method of claim 7, further comprising the step of removing skew from the horizontal and vertical lines.

10. The method of claim 7, wherein the step of determining table entries and form fields using the horizontal and vertical lines, comprises the steps of:

determining intersections between horizontal and vertical line pairs;

determining table entries from the determined intersections; and

associating horizontal lines that do not intersect other lines as part of form fields.

11. The method of claim 1, wherein the structured document is based on SGML, XML or extensions thereof.

12. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for processing electronic documents, the method steps comprising:

receiving as input an electronic document, wherein at least a portion of the electronic document is raster-based;

extracting form information from text portions and non-text portions of the electronic document; and

generating a structured document representing the extracted form information for the electronic document based on a predefined document type definition.

13. The program storage device of claim 12, wherein the instruction for extracting form information comprise instructions for performing the method steps of:

segmenting text portions and non-text portions;

separately processing the text portions and non-text portions to extract associated form information; and

combining the extracted form information of the text portions and non-text portions.

14. The program storage device of claim 12, wherein the electronic document comprises a PDF (portable document format) file.

15. The program storage device of claim 13, wherein the instructions for processing the text portions of the electronic document comprise instructions for performing the method steps of:

extracting text segments and corresponding location information of the extracted text segments; and

determining a context of each text segment using the location information.

16. The program storage device of claim 15, wherein the context of a text segment comprises one of a paragraph, heading, title and subheading.

17. The program storage device of claim 15, further comprising instructions for extracting form lines and extracting table boxes.

18. The program storage device of claim 13, wherein the instructions for processing the non-text portions comprise instructions for performing the method steps of:

extracting images and location information for the images;

converting the extracted images to black and white images;

determining horizontal and vertical lines in the images; and

determining table entries and form fields using the horizontal and vertical lines.

19. The program storage device of claim 18, wherein the instructions for determining horizontal and vertical lines comprise instructions for performing the method steps of:

performing a Hough Transformation to extract line information from an image;

determining lines in the image using the extracted line information; and

discarding lines having a slope that exceeds a predetermined threshold.

20. The program storage device of claim 18, further comprising instructions for performing the step of removing skew from the horizontal and vertical lines.

21. The program storage device of claim 18, wherein the instructions for determining table entries and form fields using the horizontal and vertical lines comprise instructions for performing the steps of:

determining intersections between horizontal and vertical line pairs;

determining table entries from the determined intersections; and

associating horizontal lines that do not intersect other lines as part of form fields.

22. The program storage device of claim 12, wherein the structured document is based on SGML, XML or extensions thereof.

23. A system for processing electronic documents, comprising:

a text differentiation module for processing an electronic document to identify text portions and non-text portions of the electronic document, wherein at least a portion of the electronic document is raster-based;

a segmentation module for segmenting the identified text portions and non-text portions;

a text processing and pattern matching module and a form information extraction module, for processing the text portions of the electronic document and extracting form information from the text portions;

an image processing and object recognition module and a form information extraction module, for processing the non-text portions of the electronic document and extracting form information from the non-text segments; and

a file generator for combining the extracted form information for the text portions and non-text portions to generate a structured document representing the extracted form information for the electronic document based on a predefined document type definition.