Method and system for creating flexible structure descriptions

• US 8,233,714 B2
• Filed: 02/02/2009
• Issued: 07/31/2012
• Est. Priority Date: 08/01/2006
• Status: Active Grant

• Current Assignee: ABBYY Development LLC
• Original Assignee: ABBYY Software
• Inventors: Zuev, Konstantin, Tuganbaev, Diar, Filimonova, Irina, Zlobin, Sergey
• Primary Examiner(s): CHAWAN, SHEELA C
• Application Number:
  US12/364,266
• Publication Number:
  US 20090175532A1
• Time in Patent Office: 1,275 Days
• Field of Search: 382/181, 382/180, 382/305, 707/999.1, 707/999, 707/999.004, 707/999.102, 707/999.003, 707/999.001, 707/999.202, 707/999.006, 707/999.002, 707/999.104, 707/999.005, 707/999.009, 707/E7.127, 707/E7.061, 707/E7.101, 707/E17.059, 707/709, 707/769, 707/736, 707/758, 375/E7.133, 375/E7.15, 375/E7.163, 348/E7.013, 348/E7.016, 709/206, 715/207, 715/234, 365/158, 365/171, 725/53, 725/135
• US Class Current: 382/181
• CPC Class Codes: G06F 40/106 Display of layout of documents; Previewing
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1452 based on positionally close symbols, e.g. amount sign or URL-specific characters
• CPC Class Codes: G06V 30/416 Extracting the logical structure, e.g. chapters, sections or page numbers; Identifying elements of the document, e.g. authors
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A method related to data capture from forms involving optical character recognition comprises detecting data fields on a scanned image; generating a flexible document description based on the detected data fields, including creating a set of search elements for each data field, each search element having associated search criteria; and training the flexible document description using a search algorithm to detect the data fields on additional training images based on the set of search elements.

Claims

1. A computer-implemented method comprising:

obtaining an image of a paper document;

detecting data fields on the image corresponding to a document type;

generating by a processor of a data capture system a flexible document description for the document type based on the detected data fields, including creating a set of search elements for each data field, each search element having one or more associated search criteria; and

training the flexible document description using a search algorithm to detect the data fields on additional training images based on the set of search elements.

2. The method of claim 1, wherein training the flexible document description comprises:

assessing a reliability of the search criteria associated with a search element to detect a given data field on the additional images; and

selectively adjusting the one or more search criteria so that the one or more search criteria are used to detect the given data field on the additional images within a defined reliability threshold.

3. The method of claim 1, further comprising receiving a user input indicating at least one of the data fields to be detected.

4. The method of claim 1, wherein the search elements comprise a type element to identify a data field by type and at least one anchor element to enable detection of a data field based on its position relative to the anchor element.

5. The method of claim 1, wherein the search elements comprise a plurality of auxiliary anchor elements, each to describe a data field.

6. The method of claim 5, wherein the auxiliary anchor elements are predefined and are selected from the group consisting of Static Text, Separator, White Gap, Barcode, Character String, Paragraph, Picture, Phone, Date, Time, Currency, Logo, Group, Table, and Repeating Item.

7. The method of claim 1, wherein the one or more associated search criteria comprises a type of the image object to detect, its physical properties, its search area, and its spatial relationships with other elements.

8. The method of claim 7, wherein the one or more associated search criteria comprise a image distortion parameter to enable handling of a distortion in images during conversion into electronic format.

9. The method of claim 8, wherein the image distortion parameter is selected from the group consisting of an indication of a quantity of optical character recognition (OCR) errors in elements of type Static Text;

an indication of a length of breaks in separator lines; and

an indication of an amount of noise objects in white spaces.

10. The method of claim 9, wherein the image distortion parameter is chosen automatically.

11. The method of claim 7, wherein the search area for an element is specified in terms of search constraints selected from the group consisting of an absolute search constraint specified in terms of a array of rectangles with specified coordinates;

a relative search constraint specified relative to the edges of the image of the paper document; and

a relative search constraint specified relative to other search elements.

12. The method of claim 11, wherein detecting the data fields comprises detecting image objects capable of distinguishing the document type from other document types.

13. The method of claim 12, wherein the image objects are selected from the group consisting of titles, and logos.

14. A non-transitory computer-readable medium having stored thereon a sequence of instructions which when executed by a system, causes the system to perform a method, comprising:

detecting data fields on an image;

generating a flexible document description based on the detected data fields, including creating a set of search elements for each data field, each search element having one or more associated search criteria; and

training the flexible document description using a search algorithm to detect the data fields on additional training images based on the set of search elements.

15. The non-transitory computer-readable medium of claim 14, wherein training the flexible document description comprises assessing a reliability of the one or more associated search criteria associated with a search element to detect a given data field on the additional images and selectively adjusting the one or more associated search criteria so that the one or more associated search criteria can be used to detect the given data field on the additional images within a defined reliability threshold.

16. The non-transitory computer-readable medium of claim 14, wherein the method further comprises receiving a user input indicating one or more of the data fields to be detected.

17. The non-transitory computer-readable medium of claim 14, wherein the search elements comprise a type element to identify a data field by type and at least one anchor element to enable detection of a data field based on its position relative to the anchor element.

18. The non-transitory computer-readable medium of claim 14, wherein the search elements comprise a plurality of auxiliary anchor elements, each to describe a data field.

19. The non-transitory computer-readable medium of claim 18, wherein the auxiliary anchor elements are predefined and are selected from the group consisting of Static Text, Separator, White Gap, Barcode, Character String, Paragraph, Picture, Phone, Date, Time, Currency, Logo, Group, Table, and Repeating Item.

20. A system comprising:

a processor; and

a memory coupled to the processor, the memory storing instructions which when executed by the processor causes the system to perform a method, comprising;

detecting data fields on an image;

generating a flexible document description based on the detected data fields, including creating a set of search elements for each data field, each search element having one or more associated search criteria; and

training the flexible document description using a search algorithm to detect the data fields on additional training images based on the set of search elements.

21. The system of claim 20, wherein training the flexible document description comprises assessing a reliability of the one or more search criteria associated with a search element to detect a given data field on the additional images and selectively adjusting the one or more of the one or more associated search criteria so that the one or more associated search criteria can be used to detect the given data field on the additional images within a defined reliability threshold.

22. The system of claim 20, wherein the method further comprises receiving a user input indicating one or more of the data fields to be detected.