Method and system for collecting data from a plurality of machine readable documents

• US 7,668,372 B2
• Filed: 08/26/2004
• Issued: 02/23/2010
• Est. Priority Date: 09/15/2003
• Status: Active Grant

• Current Assignee: Ip Ot Sub Ulc
• Original Assignee: Open Text Corporation
• Inventors: Schiehlen, Matthias
• Primary Examiner(s): Do; Anh Hong
• Application Number:
  US10/571,239
• Publication Number:
  US 20070065011A1
• Time in Patent Office: 2,007 Days
• Field of Search: 382/181, 382/161, 382/209, 382/175, 382/173, 382/157, 382/187, 382/240, 382/100, 382/228, 382/176, 706/12, 707/100, 707/10, 707/104.1, 709/203, 715/250, 715/505, 358/403
• US Class Current: 382/176
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/127 with the intervention of an operator
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields
• 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

Abstract

In a method and system for collection of data from documents present in machine-readable form, at least one already-processed document stored as a template and designated as a template document is associated with a document to be processed designated as a read document. Fields for data to be extracted are defined in the template document. Data contained in the read document are already extracted from regions that correspond to the fields in the template document. Should an error have occurred or no suitable template document having been associated given the automatic extraction of the data, the read document is shown on a screen and fields are manually inputted in the read document from which the data are extracted. After the manual input of the fields in the read document, the read document with field specifications is stored as a new template document or the previous template document is corrected corresponding to the newly input fields.

Claims

1. A method for collection of data from documents present in machine-readable form, the method performed by a computer system with a processor and memory, the method comprising the steps of:

associating at least one already-processed document stored as a template and subsequently designated as a template document with a document to be processed that is designated as a read document, fields for data to be extracted being defined in the template document, wherein associating the at least one already-processed document with the document to be processed is performed by the processor executing instructions stored in the memory;

automatically extracting data from the read document, the data contained in regions of the read document that correspond to the fields in the template document, wherein automatically extracting data is performed by the processor executing instructions stored in the memory; and

if an error occurs, or if no suitable template document is associated;

showing the read document on a screen and manually inputting fields in the read document from which the data are extracted; and

storing the read document with field specifications as a new template document, or correcting the at least one already-processed template document corresponding to the newly input fields.

2. A method according to claim 1 wherein after the manual input it is checked whether the previous field specifications contained in the template document can be corrected according to the manually input field specifications, and if this is the case a corresponding correction of the template document is executed;

the changes to the template document are then checked using the previously-processed read documents; and

in the event that said check yields that not all previously processed read documents can be correctly read with the changes, the corrections are discarded and the read document with the manually input field specifications is stored as a new template document; and

the read document with the manual field specifications is also stored as a template document when the contained field specifications of the previously used template document cannot be automatically corrected.

3. A method according to claim 1 wherein given the manual input of the fields, these are linked with a databank in which the extracted data are stored.

4. A method according to claim 3 wherein the linking occurs via input of a field in the read document and subsequent actuation of an element of the databank.

5. A method according to claim 1 wherein syntax or semantics of individual fields is evaluated upon association of a template document with a read document, and the association is executed dependent on said evaluation.

6. A method according to claim 1 wherein a plurality of read documents are evaluated as to whether surroundings are constant with regard to a field in the template document, and when this is the case this is stored and considered as a further criterion in mapping of a field of the template document to the read document.

7. A method according to claim 6 wherein in the association of the template document with the read document, a statistical evaluation is executed that associates a familiarity value with each field of the template document;

said familiarity value describing how well data of the read documents corresponding to said field could be read; and

given the evaluation of the associability of the template document with the read document the fields with a high familiarity value are rated higher than the fields with a lower familiarity value.

8. A method according to claim 1 wherein the template document is shown on the screen after the manual input of fields, and fields in the template document are manually corrected.

9. A method for collection of data from documents present in machine-readable form, the method performed by a computer system with a processor and memory, the method comprising the steps of:

associating at least one already processed document stored as a template and designated as a template document with a document to be processed designated as a read document, where fields for data to be extracted are defined in the template document;

said associating occurring by a cost function with which a similarity between the read document and template documents is calculated and a template document with a best similarity is associated with the read document; and

automatically extracting data contained in the read document from regions that correspond to the fields in the template document, wherein automatically extracting data is performed by the processor executing instructions that are stored in the memory.

10. A computer system for collecting data from documents present in machine-readable form, the computer system having a processor and a memory, the memory containing computer readable instructions which, when executed by the processor, perform the following steps:

compare at least one already processed document that is stored as a template document and designated as template document with a document to be processed designated as a read document, fields for data to be extracted being defined in the template document;

determine a similarity between the read document and the template document using a cost function,

associate a template document having a best similarity with the read document; and

automatically extract data contained in the read document from regions that correspond to the fields in the template document.

11. A computer system according to claim 10 further comprising a scanner for acquisition of documents.

12. A method of transforming a template document using a read document, the method including:

associating a stored template document with a read document, wherein the stored template document includes at least one field, and wherein the read document includes data;

automatically extracting at least some of the data from the read document, wherein the extracted data is located in at least one region of the read document that correspond to the at least one field of the stored template document;

displaying the read document on a display screen;

receiving an input that associates a region of the read document with a chosen field; and

generating a template document, wherein;

the template document is a new template document and is transformed to incorporate the chosen field associated with the read document, or

the template document is the stored template document and is transformed to incorporate the chosen field associated with the read document.

13. The method according to claim 12 wherein generating the template document further includes:

verifying that the chosen field corresponds to a proper region of a previously processed read document, wherein the read document is stored as the new template document incorporating the chosen field if;

the chosen field does not correspond to the proper region of the previously processed read document, or

a field of the stored template document cannot be automatically corrected with the chosen field.

14. A method according to claim 12 wherein the chosen field is linked with a databank in which data automatically extracted from the read document are stored.

15. A method according to claim 14 wherein the chosen field is linked via input of the chosen field in the read document and selection of an element of the databank.

16. A method according to claim 12 further comprising evaluating a syntax or semantics of the chosen field when the input is received to associate the chosen field with the region of the read document, wherein the association is executed dependent on said evaluation.

17. A method according to claim 12 further comprising evaluating a plurality of read documents to determine whether surroundings in the plurality of read documents are constant with regard to at least one field in the stored template document, wherein this evaluation is considered in mapping a field of the template document to a region of a read document.

18. A method according to claim 17 wherein associating the stored template document with the read document includes executing a statistical evaluation associating a familiarity value with the at least one field in the stored template document, the familiarity value describing how well data of the read document corresponding to the at least one field could be read.

19. A method according to claim 12 wherein generating the template document includes displaying the stored template document on a display screen and receiving an input to incorporate the chosen field into the stored template document.

20. A method of generating a template document using a read document, the method performed by a computer system having a processor and memory, the method including:

receiving a read document, wherein the read document is received from a scanner;

comparing a stored template document with the read document, wherein the stored template document includes at least one field, wherein comparing the stored template document with the read document is performed by the processor executing instructions that are stored in the memory;

displaying the read document on a display screen;

receiving an input that associates a region of the read document with a chosen field; and

incorporating the chosen field into the stored template document or creating a new template document with the read document.