Method and means for enhancing optical character recognition of printed documents

• US 5,748,807 A
• Filed: 10/15/1993
• Issued: 05/05/1998
• Est. Priority Date: 10/09/1992
• Status: Expired due to Fees

• Current Assignee: Matsushita Electric Corporation Of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Technologies, Inc. (Panasonic Holdings Corporation)
• Inventors: Lopresti, Daniel P., Sandberg, Jonathan S.
• Primary Examiner(s): Mancuso, Joseph
• Application Number:
  US08/138,467
• Time in Patent Office: 1,663 Days
• Field of Search: 382/57, 382/61, 382/309-311, 382/312, 382/317, 358/403, 358/405, 364/737, 371/53, 371/52, 371/57.2
• US Class Current: 382/310
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/12 Detection or correction of errors, e.g. by rescanning the pattern
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields

Abstract

A document marker, including first values dependent upon the layout and the contents of the document and assigned by generating or preprocessing software, is provided in machine-readable symbology on the face of a printed version of the document. The marker may include encoded document layout information and values assigned on sequences of the original text, including text-dependent decimation sequences, error correction codes or check-sums. Upon optical character recognition scanning, or other digitizing reproduction, the marker is also scanned. The scanning computer, having corresponding software, assigns second values dependent upon the layout and contents of the reproduced document. Upon comparison of the first and second decimation sequences, line and character errors can be detected and some errors corrected, thereby generating re-aligned candidate sequences. Optional error correction codes can provide further correcting capabilities, as applied to the re-aligned reproduced document sequences; and, an optional check-sum comparison can be utilized to verify the accuracy of the reproduced sequences are correct.

Claims

1. A method for permitting machine-represented characters of a machine-represented document to be more accurately recovered when scanning a printed version of said document, comprising the steps of:

assigning a first binary value to each of a plurality of original machine-represented characters in said document;

printing said first binary values in machine-readable symbology and human recognizable characters corresponding to said original machine-represented characters in a printed version of said document;

scanning said printed version to recover machine-represented characters corresponding to said printed human-recognizable characters and to recover said first binary values;

assigning a second binary value to each of said recovered machine-represented characters; and

comparing said recovered first binary values to said second binary values to identify errors in said recovered machine-represented characters.

2. The method of claim 1 further comprising the step of locating an error in respective ones of said recovered machine-represented characters based on detecting a difference between said recovered first binary values and said second binary values corresponding to respective ones of said recovered machine-represented characters.

3. The method of claim 2 further comprising the step of automatically correcting said located errors.

4. The method of claim 3 wherein said step of correcting comprises realigning said plurality of second binary values to said plurality of recovered first binary values to eliminate said errors.

5. The method of claim 4 wherein said step of correcting comprises altering said recovered machine-represented characters to generate intermediate characters adapted to be corrected to more accurately recover said original machine-represented characters.

6. The method claim 5 further comprising applying said encoded error correction information to perform at least one of error detection or correction on said intermediate characters.

7. The method of claim 1 wherein said step of assigning further comprises assigning at least one data-dependent value to said original machine-represented characters and said printing further comprises printing said data-dependent value in said printed version.

8. The method of claim 7 wherein said assigning of said data-dependent value comprises encoding error correction information for said original machine-represented characters.

9. The method of claim 7 wherein said assigning of said data-dependent value comprises calculating at least one first check-sum for said original machine-represented characters.

10. The method of claim 7 wherein said assigning of said data-dependent value comprises calculating at least one first check-sum for said original machine-represented characters, and said method further comprises the steps of:

calculating a second check-sum for said intermediate characters;

comparing said first check-sum to said second check-sum; and

detecting the existence of at least one error in said intermediate characters when said check-sums differ.

11. The method of claim 1 wherein said step of comparing identifies insertion errors.

12. The method of claim 1 wherein said step of comparing identifies deletion errors.

13. A method for producing more accurate optical character recognition reproductions of documents having lines of original printed text and at least one document marker including a plurality of first decimation sequences, each corresponding to a sequence of said printed text, and at least one first check-sum value calculated on said sequence of printed text, comprising the steps of:

creating an electronic document comprising a plurality of first reproduced text sequences by optically scanning said original printed text;

optically scanning said at least one document marker;

decoding said plurality of first decimation sequences from said scanned document marker;

decimating said reproduced text into a plurality of second decimation sequences;

calculating the edit distances between corresponding lines of original printed text and reproduced printed text by comparing said first and said second decimation sequences;

comparing the edit distances and identifying line insertion and deletion errors in said reproduced text when said edit distances differ by more than a predetermined amount;

correcting detected line insertion and deletion errors;

comparing each of said corresponding plurality of first and second decimation sequences;

identifying text errors in said reproduced text at the sequence location at which said decimation sequences differ;

substituting different characters in said sequence locations at which text errors have been identified to produce at least one second reproduced text sequence;

calculating a second check-sum for each of said at least one second reproduced text sequences;

comparing said second check-sum to said first check-sum; and

verifying the accuracy of said second reproduced text sequence when said first and said second check-sums are equal.

14. The method of claim 13 wherein said document marker further comprises a plurality of error correction codes for said original text and further comprising the step of applying said corresponding error correction code to said at least one second reproduced text sequence prior to calculating said second check-sum for said text.