System and method for creating a searchable word index of a scanned document including multiple interpretations of a word at a given document location
View Patent ↗Multiple recognition engines ( 110 ) provide different interpretations ( 116 ) of a word at a given location within a scanned document ( 108 ). A word node corresponding to each unique interpretation is stored within a word index ( 102 ), with each word node being linked to word nodes of previously and subsequently recognized words.
1. A method in a computer system for creating a searchable word index of a scanned document, the method comprising:
generating a first interpretation of a word at a given location within the scanned document using a first recognition engine;
generating a second interpretation of the word using a second recognition engine, wherein the second interpretation is different from the first interpretation;
storing a first word node in the searchable word index associating the first interpretation of the word and the location of the word within the scanned document;
storing a second word node in the searchable word index associating the second interpretation of the word and the location of the word within the scanned document; and
linking the first and second word nodes to at least one word node of an adjacent word from the scanned document.
2. The method of claim 1 , wherein the first and second recognition engines employ different optical character recognition (OCR) techniques.
3. The method of claim 1 , wherein the location of the word is defined by a bounding rectangle.
4. The method of claim 3 , wherein the bounding rectangle is defined by at least two coordinates, each coordinate comprising a percentage of a width and a height of the scanned document.
5. The method of claim 1 , wherein the adjacent word comprises a subsequent word in the scanned document.
6. The method of claim 5 , wherein the first and second word nodes are each bidirectionally linked to adjacent words from the scanned document on opposite sides of the word.
7. The method of claim 1 , further comprising:
generating a third interpretation of the word using a third recognition engine;
determining whether the third interpretation of the word is contained within a word list; and
storing a third word node in the searchable word index when the third interpretation of the word is contained within the word list, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
8. The method of claim 7 , wherein the word list comprises a dictionary.
9. The method of claim 1 , further comprising:
generating a third interpretation of the word using a third recognition engine;
determining whether the third interpretation of the word contains an improbable character triplet;
storing a third word node in the searchable word index when the third interpretation of the word does not contain an improbable character triplet, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
10. The method of claim 9 , wherein an improbable character triplet comprises three consecutive characters not found within a word of a dictionary.
11. A system for creating a searchable word index of a scanned document, the system comprising:
a first recognition engine stored on a computer readable medium and configured to generate a first interpretation of a word at a given location within the scanned document;
a second recognition engine stored on a computer readable medium and configured to generate a second interpretation of the word, wherein the second interpretation is different from the first interpretation;
an index creation component stored on a computer readable medium and configured to store first and second word nodes in the searchable word index, the first word node associating the first interpretation of the word and the location of the word within the scanned document and the second word node associating the second interpretation of the word and the location of the world within the scanned document; and
a linking component stored on a computer readable medium and configured to link the first and second word nodes to a word node of an adjacent word from the scanned document.
12. The system of claim 11 , wherein the first and second recognition engines employ different optical character recognition (OCR) techniques.
13. The system of claim 11 , wherein the location of the word is defined by a bounding rectangle.
14. The system of claim 13 , wherein the bounding rectangle is defined by at least two coordinates, each coordinate comprising a percentage of a width and a height of the scanned document.
15. The system of claim 11 , wherein the linking component is configured to bidirectionally link the first and second word nodes to adjacent words from the scanned document on opposite sides of the word.
16. The system of claim 11 , further comprising:
a third recognition engine configured to generate a third interpretation of the word using a third recognition engine; and
a word filter configured to determine whether the third interpretation of the word is contained within a word list;
wherein the index creation component is further configured to store a third word node in the searchable word index when the third interpretation of the word is contained within the word list, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
17. The system of claim 16 , wherein the word list comprises a dictionary.
18. The system of claim 11 , further comprising:
a third recognition engine configured to generate a third interpretation of the word using a third recognition engine; and
a word filter configured to determine whether the third interpretation of the word contains an improbable character triplet;
wherein the index creation component is further configured to store a third word node in the searchable word index when the third interpretation of the word does not contain an improbable character triplet, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
19. The system of claim 18 , wherein an improbable character triplet comprises at three consecutive characters not found within a word of a dictionary.
20. A computer program product on a computer-readable medium for creating a searchable word index of a scanned document, the computer program product comprising:
program code for generating a first interpretation of a word at a given location within the scanned document using a first recognition engine;
program code for generating a second interpretation of the word using a second recognition engine, wherein the second interpretation is different from the first interpretation;
program code for storing a first word node in the searchable word index associating the first interpretation of the word and the location of the word within the scanned document;
program code for storing a second word node in the searchable word index associating the second interpretation of the word and the location of the word within the scanned document; and
program code for linking the first and second word nodes to at least one word node of an adjacent word from the scanned document.
21. The computer program product of claim 20 , wherein the first and second recognition engines employ different optical character recognition (OCR) techniques.
22. The computer program product of claim 20 , wherein the location of the word is defined by a bounding rectangle.
23. The computer program product of claim 22 , wherein the bounding rectangle is defined by at least two coordinates, each coordinate comprising a percentage of a width and a height of the scanned document.
24. The computer program product of claim 20 , wherein the adjacent word comprises a subsequent word in the scanned document.
25. The computer program product of claim 24 , wherein the program code for linking the first and second word nodes is configured to bidirectionally link the first and second word nodes to adjacent words from the scanned document on opposite sides of the word.
26. The computer program product of claim 20 , further comprising:
program code for generating a third interpretation of the word using a third recognition engine;
program code for determining whether the third interpretation of the word is contained within a word list; and
program code for storing a third word node in the searchable word index when the third interpretation of the word is contained with the word list, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
27. The computer program product of claim 26 , wherein the word list comprises a dictionary.
28. The computer program product of claim 20 , further comprising:
program code for generating a third interpretation of the word using a third recognition engine;
program code for determining whether the third interpretation of the word contains an improbable character triplet; and
program code for storing a third word node in the searchable word index when the third interpretation of the word does not contain an improbable character triplet, the third word node associating the third interpretation of the word and the location of the word within the scanned document.
29. The computer program product of claim 28 , wherein an improbable character triplet comprises three consecutive characters not found within a word of a dictionary.
30. The method of claim 1 , wherein both the first and the second interpretations of the word are searchable in the word index and may be identified by performing a search of the word index.