Redigitization system and service

• US 9,330,323 B2
• Filed: 04/29/2012
• Issued: 05/03/2016
• Est. Priority Date: 04/29/2012
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Simske, Steven J, Liu, Samson J.
• Primary Examiner(s): Wu, Jingge
• Application Number:
  US14/364,743
• Publication Number:
  US 20150049949A1
• Time in Patent Office: 1,465 Days
• US Class Current: 1/1
• CPC Class Codes: G06V 10/98 Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns
• CPC Class Codes: G06V 30/224 of printed characters having additional code marks or containing code marks

Abstract

A system and method to error correct extant electronic documents is disclosed. An electronic document may be rasterized to obtain a pixel representation of the electronic document (e.g., raster image). One or more optical character recognition (OCR) tasks may be performed on the raster image of the electronic document. Errors discovered by the OCR tasks may be corrected and a customized error corrected version of the electronic document may be created and stored. If the author of the electronic document is known, the raster image may be compared to a personalized tf*idf error dictionary associated with the author to determine known OCR errors specific to the author. The raster image may also be compared to a personalized electronic error dictionary associated with the author to determine known typographical errors specific to the author.

Claims

1. A method comprising:

rasterizing an electronic document to obtain a raster image of the electronic document;

determining the author of the electronic document;

performing one or more optical character recognition (OCR) tasks on the raster image of the electronic document, performing the OCR tasks including identifying digitization errors in the electronic document based on a comparison to a personalized tf*idf error dictionary associated with the author to determine known OCR errors specific to the author, the personalized tf*idf error dictionary representing (i) whether a term occurred or not, (ii) how many times the term occurred, (iii) what percent of words are each term, (iv) using log instead of linear scales for the number of occurrences of the term, (v) or a combination thereof;

correcting errors discovered by the OCR tasks; and

creating a customized error corrected version of the electronic document.

2. The method of claim 1, further comprising storing the customized error corrected version of the electronic document.

3. The method of claim 1, at least one of the OCR tasks comprising:

comparing the raster image to an OCR error dictionary to determine known OCR errors.

4. The method of claim 1, further comprising:

searching for multiple versions of the same electronic document;

comparing the differential error rates among the multiple versions of the electronic document;

creating a master error corrected custom redigitized electronic document; and

storing the master error corrected custom redigitized electronic document.

13. The method of claim 1, further comprising determining if the electronic document originated as an electronic document or a scanned version of a paper document.

5. An apparatus comprising:

a processor circuit;

a processing engine to access stored electronic documents;

an optical character recognition (OCR) engine under control of the processor;

a rasterizer to create raster images of the electronic documents;

an OCR error dictionary stored in a memory;

a personalized tf*idf dictionary stored in the memory; and

a personalized electronic error dictionary stored in the memory,

the OCR engine to perform OCR tasks on the raster images using the OCR error dictionary, the personalized tf*idf dictionary, and the personalized electronic error dictionary, the OCR tasks to;

determine errors in the raster images;

correct the errors in the raster images;

create a customized error-corrected version of the electronic document; and

store the customized error-corrected version of the electronic document in the memory.

6. An article of manufacture comprising a non-transitory computer-readable storage medium containing instructions that if executed enable a system to:

access an electronic document;

obtain a pixel representation of a first version of an electronic document;

perform one or more optical character recognition (OCR) tasks on the pixel representation of the electronic document;

correct errors in the first version of the electronic document discovered by the OCR tasks;

search for a second version of the electronic document; and

create a customized error corrected version of the electronic document based on a comparison of a differential error rate between the first version of the electronic document and the second version of the electronic document.

7. The article of claim 6, further comprising instructions that if executed enable the system to store the customized error corrected version of the electronic document.

8. The article of claim 6, further comprising instructions that if executed enable the system to compare the pixel representation of the electronic document to an OCR error dictionary to determine known OCR errors.

9. The article of claim 6, further comprising instructions that if executed enable the system to determine the author of the electronic document.

10. The article of claim 9, further comprising instructions that if executed enable the system to compare the pixel representation of the electronic document to a personalized tf*idf error dictionary associated with the author to determine known OCR errors specific to the author, the personalized tf*idf error dictionary representing one or more of (i) whether a term occurred or not, (ii) how many times the term occurred, (iii) what percent of words are each term and (iv) using log instead of linear scales for the number of occurrences of the term.

11. The article of claim 9, further comprising instructions that if executed enable the system to compare the pixel representation of the electronic document to a personalized electronic error dictionary associated with the author to determine known typographical errors specific to the author.

12. The article of claim 9, further comprising instructions that if executed enable the system to search for the second version of the electronic document based on a similarity level to the first version of the electronic document.