Processing an electronic document for information extraction

• US 7,251,644 B2
• Filed: 08/02/2004
• Issued: 07/31/2007
• Est. Priority Date: 12/04/2003
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Viola, Paul, Rinker, James, Law, Hiu Chung
• Primary Examiner(s): Mofiz; Apu
• Assistant Examiner(s): Nguyen; Cindy
• Application Number:
  US10/909,534
• Publication Number:
  US 20050125746A1
• Time in Patent Office: 1,093 Days
• Field of Search: 707/1, 707/5
• US Class Current: 1/1
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

The present invention relates to a method of automatically processing an electronic document for routing over a computer network. The method includes recognizing text in the document to identify a candidate address, accessing a collection of potential destinations and comparing the candidate address to the collection of potential destinations to determine a destination for the document.

Claims

1. A method of automatically processing an electronic document for routing over a computer network, comprising:

recognizing text in the document to identify a candidate address indicative of a portion of the document containing a potential destination;

accessing a collection of potential destinations;

computing a first weighted score and a contiguous weighted score for at least one potential destination, wherein the first weighted score is indicative of a difference between a word in the candidate address and a word associated with the at least one potential destination, and wherein the contiguous weighted score is indicative of a difference between contiguous words in the candidate address and words associated with the at least one potential destination;

computing a combined score for the at least one potential destination based on the first weighted score and the contiguous weighted score;

determining a recipient from the collection of potential destinations based on the combined score; and

routing the document to the recipient.

2. The method of claim 1 and further comprising:

identifying features associated with the candidate addresses.

3. The method of claim 2 and further comprising:

assigning a score to the candidate address based on the features of the candidate address.

4. The method of claim 2 wherein the features relate to at least one of a location of the candidate address, a distance from the candidate address to a selected word and text of the candidate address.

5. The method of claim 1 and further comprising:

computing a value indicative of a difference between the candidate address and at least one potential destination.

6. The method of claim 5 wherein the value is computed based on the difference between a word in the candidate address and a word associated with the at least one potential destination.

7. The method of claim 1 wherein the collection of potential destinations is stored in a database with associated fields including at least a first name, a last name and an email address of a recipient.

8. The method of claim 1 wherein the document is a fax.

9. A computer storage medium having instructions that, when implemented on a computer, cause the computer to process a document for routing to a selected address, the instructions comprising:

a recognition module adapted to recognize words within the document, the recognized words being indicative of a portion of the document containing containing a potential destination;

an identification module adapted to identify features of the recognized words within the document and assign a relevance score to each recognized word based on the features, wherein the features relate to at least one of a location of a word in the text, a distance from a word in the text to a nearby word in the text and text of a word, and wherein the identification module is further adapted to select relevant word candidates from the recognized words based on the relevance score;

a comparison module adapted to compare the relevant word candidates to words associated with a collection of addresses and determine a recipient based on the text and the features; and

a routing module adapted to route the document to the recipient.

10. The computer storage medium of claim 9 wherein the recognition module is adapted to perform optical character recognition.

11. The computer storage medium of claim 9 wherein the words associated with the collection of addresses correspond to at least one of a first name, a last name and an email address.

12. The computer storage medium of claim 9 wherein the features relate to at least one of a location of a word in the text, a distance from a word in the text to a nearby word in the text and text of a word.

13. A method of processing an electronic document for routing over a computer network, comprising:

recognizing text in the document to identify a plurality of candidate addresses, wherein each of the plurality of candidate addresses is indicative of a portion of the document containing a potential destination;

accessing a collection of potential destinations;

identifying features associated with the plurality of candidate addresses;

computing weighted relevance scores for the plurality of candidate addresses, wherein the scores are weighted based on the identified features;

generating a list of the candidate addresses;

sorting the list of candidate addresses based on the weighted relevance scores;

determining a recipient from the collection of potential destinations based on the list; and

routing the document to the recipient.

14. The method of claim 13 wherein identifying features further comprises identifying features relating to at least one of a location of the candidate address, a distance from the candidate address to a selected word and text of the candidate address.

15. The method of claim 13 wherein identifying features further comprises identifying a substring indicative of one of an email address and an alias.

16. The method of claim 13 wherein the collection of potential destinations is stored in a database with associated fields including at least a first name, a last name and an email address of a recipient.

17. The method of claim 13 wherein the document is a fax.

18. The method of claim 13 and further comprising generating the list of candidate addresses based on differences between words in the plurality of candidate addresses and words in the collection of potential destinations.

19. The method of claim 13 wherein determining a recipient comprises identifying a difference between a word in a candidate address and a word in the collection of potential destinations.