Cryptographic analysis system

• US 4,610,025 A
• Filed: 06/22/1984
• Issued: 09/02/1986
• Est. Priority Date: 06/22/1984
• Status: Expired due to Fees

• Current Assignee: Devshale Oil & Gas Corporation
• Original Assignee: Champollion Group
• Inventors: Blum, Eric, Pierce, Wilbur
• Primary Examiner(s): NOT, DEFINED
• Application Number:
  US06/623,739
• Time in Patent Office: 802 Days
• Field of Search: 382/9, 382/30, 382/40, 382/48, 382/39, 382/34, 382/36, 382/61, 358/263
• US Class Current: 382/177
• CPC Class Codes: G06K 19/06037 multi-dimensional coding
• CPC Class Codes: G06K 7/143 Glyph-codes
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/262 using context analysis, e.g. lexical, syntactic or semantic context

Abstract

A system for recognizing the content of a communication in symbolic language and composed of plural glyphs arranged in a predetermined order, each glyph being the smallest (lowest) informational unit of the language. The system includes a device for inputting a stream of data indicative of the plural glyphs, such as formed in a page of text. That stream is input into a storage means. The stored data is horizontally segmented into discrete lines of text and is then vertically segmented into individual glyphs. Each individual glyph is assigned a unique identifier, whereby all substantially identical glyphs are represented by the same identifier. The identifiers are arranged in a sequence corresponding to the sequence in which the glyphs appeared in the communication, thus representing glyph "words". The system then applies decryption routines which include general cryptographic techniques to the identifiers, their sequences and their interrelationships to determine the equivalent symbol of language corresponding to each identifier. Once the symbol of language corresponding to each identifier has been determined, the machine code equivalent (i.e., code capable of being "understood" and utilized by an electronic computer) is substituted for each identifier, so as to provide a machine readable code representation of the communication, e.g., page of text.

Claims

1. A system for recognizing the content of a communication in symbolic language having rules, said communication comprising a plurality of glyphs arranged in a predetermined order and defining plural glyph words, each of said glyphs being a discrete element generally comprising the smallest meaningful informational unit of said language, said system comprising sensory input means for sensing said plurality of glyphs and inputting a stream of data indicative thereof into storage menas, separating means for separating said data into a plurality of glyphs, compiling means for assigning a unique identifier to each set of substantially identical glyphs, pattern grouping means for arranging said identifiers in a grouped arrangement corresponding to the arrangement of said glyph words in said communication, decryption means including language and dictionary storage means for applying general cryptographic techniques to said identifiers to analyze said grouped identifiers in terms of contextual patterns by their sequences within at least one word and their interrelationships as words, to thereby determine the equivalent symbol of language corresponding to each of said identifiers.

2. The system of claim 1, further comprising assembling means, said assembling means arranging said symbols of language in an arrangement corresponding to the arrangement of said glyphs in said communication to replicate the informational content of said communication.

3. The system of claim 2, further comprising substituting means, said substituting means substituting computer recognizable code for each of said symbols of language making up said communication for use of said communication by computer means.

4. The system of claim 3, wherein said communication comprises a document.

5. The system of claim 4, wherein said sensory input means comprises means for providing a bit-mapped representation of the glyphs appearing in said document.

6. The system of claim 5, wherein said sensory input means comprises means for producing a binary representation of the glyphs appearing on said document.

7. The system of claim 6, wherein said sensory input means comprises a facsimile device.

8. The system of claim 5, wherein said separating means comprises horizontal segmentation means and vertical segmentation means, wherein said horizontal segmentation means breaks said bit-mapped representation of said document into individual lines of text and said vertical segmentation means breaks each of said lines of text into discrete glyphs.

9. The system of claim 8, wherein said separating means includes means for breaking ligatures into their respective component symbols of language.

10. The system of claim 1, wherein said compiling means identifies various physical characteristics of each of said glyphs, to determine which of said glyphs are substantially identical to other glyphs.

11. The system of claim 10, further comprising counting means for determining the number of glyphs contained in each set of substantially identical glyphs.

12. The system of claim 11, wherein said pattern grouping means arranges said identifiers in a plurality of individual sequences, wherein each individual sequence corresponds to a particular word in said language.

13. The system of claim 1, wherein said communication is a document and further comprising document type analysis means, said document type analysis means analyzing glyph patterns and the arrangement of groups of glyphs in said document to determine and classify said document as being that of a particular type.

14. The system of claim 13, wherein said decryption means selects an appropriate set of decryption routines based on classification information supplied by the document type analysis means.

15. The system of claim 14, wherein said decryption means decodes a document, which has been determined by said document type analysis means to contain general textual information, through the use of means for analyzing short words, means for matching a limited dictionary and means for matching a full dictionary.

16. The system of claim 14, wherein said decryption means decodes a document, which has been determined by the document type analysis means to be a business letter, through the use of look-up means based on geography specific keywords, means for analyzing short words, means for matching a limited dictionary pattern and means for matching a full dictionary.

17. The system of claim 14, wherein said decryption means decodes a document, which has been determined by said document type analysis means to be a financial report, through the use of look-up means based on geography specific keywords, means for matching a limited dictionary and cryptarythm analysis means.

18. The system of claim 14, wherein said decryption means includes inconsistency determining means which determines whether the results produced by each of said decryption routines is consistent with the glyph/symbol correspondences previously determined during other decryption routines.

19. The system of claim 18, wherein said inconsistency determining means analyzes said inconsistencies, removes certain previously established glyph/symbol correspondences and then selects an appropriate decryption routine to be repeated based on said analysis.

20. The system of claim 18, further comprising ambiguity sensing means, whereupon whenever said inconsistency determining means determines that there are no inconsistencies, said ambiguity sensing means determines whether all of said glyphs have been identified and, if not, selects the next decryption routine to be performed.

21. The system of claim 13, wherein said cryptographic means comprises look-up means based on geography specific keywords, said look-up means analyzing the document to determine whether certain preselected glyph patterns appear at certain predetermined locations on said document.

22. The system of claim 1, wherein said decryption means comprises a plurality of analytical means for identifying correspondences between glyphs and symbols of language, without relying to a substantial degree on the shape or geometry of the respective glyphs.

23. The system of claim 1, wherein said decryption means comprises pattern matching means which compares glyph patterns in glyph words with letter patterns in respective keywords selected from a list of keywords.

24. The system of claim 23, wherein said list of keywords includes a limited number of frequently used words.

25. The system of claim 23, wherein said list of keywords comprises a general dictionary of said language.

26. The system of claim 1, wherein said decryption means includes means for analyzing short words for determining correspondences between glyphs and symbols of language, by analyzing interrelationships among glyphs in glyph words containing fewer than five glyph letters.

27. The system of claim 26, wherein the means for analyzing short words utilizes and compares information pertaining to the frequency with which particular glyphs and glyph patterns occur, the frequency at which certain characters and groups of characters occur, respectively, in said language.

28. The system of claim 1, wherein said decryption means further comprises cryptarithm means for identifying those glyphs which correspond to symbols of language comprising digits, said cryptarithm means analyzing the context in which those glyphs occur, including their interrelationship with other glyphs which correspond to digits.

29. The system of claim 28, wherein said cryptarithm means comprises means for establishing restriction sets which limit the possible digit values for those glyphs corresponding to digits.

30. The system of claim 29, wherein said restriction sets are established by eliminating certain glyphs from consideration for being the digit zero.

31. The system of claim 29, wherein said restriction sets are established through the application of arithmetic rules and manipulations to said glyphs corresponding to digits.

32. A method of mechanized recognition of the content of a communication in a symbolic language having rules, said communication comprising a plurality of glyphs arranged in a predetermined order and defining plural glyph words, each of said glyphs being a discrete element generally comprising the smallest meaningful informational unit of said language, wherein said method comprises the steps of storing language and dictionary information for later use, sensing said plurality of glyphs and inputting a stream of data indicative thereof into storage means, separating said data into a plurality of said glyphs, assigning a unique identifier to each set of substantially identical glyphs, arranging said identifiers in a grouped arrangement corresponding to the arrangement of said glyph words in communication and decrypting said identifiers based on said stored language and dictionary information by applying general cryptographic techniques to said identifiers to analyze said grouped identifiers in terms of contextual patterns by their sequences within at least one word and their interrelationships as words, to determine the equivalent symbol of language corresponding to each of said identifiers.

33. The method of claim 32, further comprising the step of arranging said symbols of language in an arrangement corresponding to the arrangement of said glyphs in said communication to replicate the informational content of said communication.

34. The method of claim 33, further comprising the step of substituting computer recognizable code for each of said symbols of language making up said communication for use of said communication by computer means.

35. The method of claim 34, wherein said communication comprises a document and wherein said step of sensing and inputting a stream of data indicative of the glyphs of said communication is accomplished by converting said document into a bit-mapped representation thereof.

36. The method of claim 35, wherein said step of separating said symbolic language into a plurality of glyphs is accomplished by horizonally segmenting said bit-mapped representation into horizontal lines of text and then vertically segmenting each of said lines into said glyph elements.

37. The method of claim 36, wherein the step of assigning unique identifiers to each set of substantially identical glyphs is achieved by identifying various physical characteristics of each of said glyphs and then determining which of said glyphs are substantially identical to other of said glyphs.

38. The method of claim 37, wherein said pattern grouping means further comprises the step of arranging said identifiers into a plurality of individual sequences, wherein each of said individual sequences corresponds to a particular word in said language.

39. The method of claim 35, further comprising the step of analyzing said document including the arrangement of glyphs and groups of glyphs on said document, to determine and classify said document as being that of a particular type.

40. The method of claim 39, wherein the step of decrypting said identifiers includes the step of selecting an appropriate set of decryption routines based on said document type classification.

41. The method of claim 40, wherein the decrypting step includes performing of short word analysis, limited dictionary pattern matching and full dictionary pattern matching, where said document has been determined to contain general textual information.

42. The method of claim 40, wherein said decrypting step includes performing geography specific keyword look-ups, short word analysis, limited dictionary pattern matching and full dictionary pattern matching, where said document has been determined to be a business letter.

43. The method of claim 40, wherein said decrypting step includes performing geography specific keyword look-ups, limited dictionary pattern matching and cryptarithm analysis, where said document has been determined to be a financial report.

44. The method of claim 35, wherein the step of decrypting comprises identification of correspondences between glyphs and symbols of language, without relying to a substantial degree on the shape or geometry of the respective glyphs.

45. The method of claim 35, wherein the decrypting step comprises the process of comparing glyph patterns in glyph words with letter patterns in a selected list of keywords.

46. The method of claim 35, wherein the decrypting step comprises the step of analyzing short words to determine correspondences between glyphs and symbols of language contained in words having fewer that five glyph letters.

47. The method of claim 35, wherein the decrypting step comprises the step of performing geography specific keyword look-ups on said page to determine whether certain pre-selected glyph patterns appear at certain predetermined locations on said page.

48. The method of claim 35, wherein the decrypting step comprises the step of identifying those glyphs which correspond to digits and analyzing the context in which those glyphs occur, including their interrelationships with glyphs to determine their digit value.

49. The method of claim 48, wherein said interrelationships include mathematical relationships.