System and method for disambiguating phonetic input

• US 7,395,203 B2
• Filed: 07/30/2003
• Issued: 07/01/2008
• Est. Priority Date: 07/30/2003
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Tegic Communications, Inc. (Microsoft Corporation)
• Inventors: Zhang, Lu, Wong, Keng Chong, He, Lian, Meurs, Pim van, Lai, Jenny Huang-Yu, Wu, Jianchao
• Primary Examiner(s): Chawan; Vijay
• Application Number:
  US10/631,543
• Publication Number:
  US 20050027524A1
• Time in Patent Office: 1,798 Days
• Field of Search: 704/1, 704/260, 704/9, 704270-275, 704/201, 704/235, 715/535, 715/703, 400/110
• US Class Current: 704/235
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams

Abstract

A system and method for inputting Chinese characters using Pinyin without requiring the entry of a delimiter key between Pinyin entries in a reduced keyboard is disclosed. The system searches for all possible single or multiple Pinyin spellings based on the entered Latin alphabets. Once the user has completed the inputting of the Pinyin spellings for desired Chinese phrase or characters, all possible matching phrases or characters are displayed on screen and off-screen due to screen size. The user then scrolls through a list of matching phrases or characters and selects the desired one by clicking.

Claims

1. A method for disambiguating ambiguous input sequences entered by a user and generating textual output in Chinese language, said method comprising the steps of:

entering an input sequence into a user input device;

wherein said user input device comprises a plurality of input means, each of said input means being associated with a plurality of phonetic characters, an input sequence being generated each time when an input is selected by said user input device, said generated input sequence having a textual interpretation that is ambiguous due to the plurality of phonetic characters associated with said inputs, data consisting of a plurality of input sequences and, associated with each input sequence, a set of phonetic sequences whose spellings correspond to the input sequence, and a database containing a plurality of phonetic sequences and, associated with each phonetic sequence, a set of ideographic character sequences which correspond to the phonetic sequences;

comparing the input sequence with said phonetic sequence database and finding matching phonetic entries;

optionally displaying one or more matched phonetic entries;

matching said phonetic entries with said ideographic database; and

optionally displaying one or more matched ideographic characters.

2. The method of claim 1, further comprising the step of:

prioritizing phonetic sequences that match an input sequence and prioritizing ideographic sequences that match a phonetic sequence according to a linguistic model.

3. The method of claim 2, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of a radical;

alphabetically ordered;

frequency of occurrence of ideographic sequences or phonetic sequences in formal, conversational written, or conversational spoken text;

frequency of occurrence of ideographic sequences or phonetic sequences when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current input sequence entry; and

recency of use or repeated use of phonetic or ideographic sequences by the user or within an application program.

4. The method of claim 2, further comprising the step of:

once an ideographic character sequence is selected, changing the associated priority of said matching phonetic sequence and sequence of ideographic characters.

5. The method of claim 1, wherein said set of phonetic characters comprises at least one of following:

Latin alphabet;

Bopomofo alphabet also known as Zhuyin;

digits; and

punctuation.

6. The method of claim 1, wherein said phonetic sequences comprise single syllables.

7. The method of claim 1, wherein said phonetic sequences comprise single and multi-syllables.

8. The method of claim 1, wherein said phonetic sequences comprise user generated sequences.

9. The method of claim 1, wherein said phonetic syllables and said corresponding ideographic characters are stored in at least one data structure.

10. The method of claim 9, wherein the data structures are ordered by grammatical categories.

11. The method of claim 1, wherein all monosyllabic phonetic syllables are stored in a single data structure and said corresponding phonetic syllables that form a word or phrase and one or more ideographic characters that match said word or phrase are stored in at least one data structure.

12. The method of claim 1, wherein an object is added to the database if an object does not exist for an input sequence.

13. The method of claim 12, wherein in absence of matching phonetic sequences in said database, a sequence of matching phonetic sequences is automatically generated based on single and optionally multi-syllable phonetic sequences.

14. The method of claim 13, wherein said sequence of matching phonetic sequences is narrowed down through user interaction.

15. The method of claim 13, wherein a sequence of matching ideographic sequences is automatically generated based on matching phonetic sequences to ideographic sequences.

16. The method of claim 15, wherein a sequence of matching ideographic sequences is narrowed down through user interaction.

17. The method of claim 16, wherein once a selection has been made, said matching input sequence, said matching phonetic sequence and said matching ideographic sequence are added to a data structure.

18. The method of claim 12, wherein a desired phonetic sequence and corresponding ideographic character sequence are specified through a second input mechanism.

19. The method of claim 1, wherein the user can specify a particular tone for the phonetic syllable.

20. The method of claim 19, wherein one of the plurality of inputs is associated with a special wildcard input that is associated with any or all tones.

21. The method of claim 1, wherein the user can specify an explicit syllable separator.

22. The method of claim 1, further comprising the step of:

when the user enters a sequence of phonetic characters, returning a sequence of phonetic sequences of exact matches and predictions that partially match.

23. The method of claim 22, wherein said sequence of phonetic sequences is ordered according to a linguistic model.

24. The method of claim 23, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of phonetic sequences or ideographic sequences in formal or conversational written text;

frequency of occurrence of phonetic sequences or ideographic when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character sequence entry; and

recency of use or repeated use of phonetic sequences by the user or within an application program.

25. The method of claim 1, further comprising the step of:

once the user has selected a sequence of ideographic characters, presenting the user with a list of sequences of one or more ideographic characters.

26. The method of claim 25, wherein said list of sequences is ordered according to a linguistic model.

27. The method of claim 26, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of ideographic characters in formal or conversational written text;

frequency of occurrence of ideographic characters when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character entry; and

recency of use or repeated use of ideographic characters by the user or within an application program.

28. The method of claim 1, wherein the matches between said input sequence and said phonetic sequences are part of confusion sets.

29. The method of claim 28, wherein the user can select which confusion sets are active.

30. The method of claim 28, wherein one of the plurality of inputs is associated with providing alternative phonetic sequence interpretations of the input sequence based on confusion sets or misspellings.

31. The method of claim 28, wherein one of the plurality of inputs is associated with providing alternative ideographic interpretations of an input sequence, based on confusion sets or misspellings.

32. The method of claim 28, wherein the system adapts to the common misspellings or confusion sets of the user.

33. The method of claim 1, wherein the user can enter partial syllables for each of the multiple syllable words.

34. The method of claim 33, wherein the number of partial keystrokes for each syllable is one.

35. The method of claim 1, wherein the user identifies an initial sound and a final sound.

36. The method of claim 1, wherein one of the plurality of inputs is associated with a special wildcard input that is associated with zero or one of said phonetic characters.

37. The method of claim 1, where the phonetic sequences comprise matching entries in any of English, and other alphabetic languages.

38. A system for disambiguating ambiguous input sequences entered by a user and generating textual output in Chinese language, said system comprising:

a user input device having a plurality of input means, each of said input means being associated with a plurality of phonetic characters, an input sequence being generated each time when an input is selected by said user input device, said generated input sequence having a textual interpretation that is ambiguous due to the plurality of phonetic characters associated with said inputs;

a database containing a plurality of input sequences and, associated with each input sequence, a set of phonetic sequences whose spellings correspond to the input sequence;

a database containing a plurality of phonetic sequences and, associated with each phonetic sequence, a set of ideographic character sequences which correspond to the phonetic sequences;

means for comparing the input sequence with said phonetic sequence database and finding matching phonetic entries;

means for matching said phonetic entries with said ideographic database; and

an output device for displaying one or more matched phonetic entries and matched ideographic characters.

39. The system of claim 38, further comprising:

means for prioritizing phonetic sequences that match an input sequence and prioritizing ideographic sequences that match a matching phonetic sequence according to a linguistic model.

40. The system of claim 39, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of ideographic sequences or phonetic sequences in formal or conversational written text;

frequency of occurrence of ideographic sequences or phonetic sequences when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current input sequence entry; and

recency of use or repeated use of phonetic or ideographic sequences by the user or within an application program.

41. The system of claim 39, further comprising:

means for changing the associated priority of the matching phonetic sequence, and sequence of ideographic characters once an ideographic character sequence is selected.

42. The system of claim 39, wherein the matches between the input sequence and the phonetic sequences are part of confusion sets.

43. The system of claim 42, wherein the user can select which confusion sets are active.

44. The system of claim 43, wherein one of the plurality of inputs is associated with providing alternative phonetic sequence interpretations of the input sequence based on confusion sets or misspellings.

45. The system of claim 42, wherein the system adapts to the common misspellings or confusion sets of the user.

46. The system of claim 38, wherein said set of phonetic characters comprises the Latin alphabet.

47. The system of claim 38, wherein said set of phonetic characters comprises the Bopomofo alphabet also known as Zhuyin.

48. The system of claim 47, wherein once a selection has been made the matching input sequence, the matching phonetic sequence and the matching ideographic sequence is added to memory.

49. The system of claim 38, wherein said phonetic sequences comprise single syllables.

50. The system of claim 38, wherein said phonetic sequences comprise both single and multi-syllables.

51. The system of claim 38, wherein said phonetic sequences comprise user generated sequences.

52. The system of claim 38, wherein said phonetic syllables and said corresponding ideographic characters are stored in a single tree.

53. The system of claim 38, wherein all monosyllabic phonetic syllables are stored in a single tree and the corresponding phonetic syllables that form a word or phrase and one or more ideographic characters that match said word or phrase are stored in a single tree.

54. The system of claim 38, wherein an object is added to a custom database if an object does not exist for an input sequence.

55. The system of claim 54, wherein in absence of matching phonetic sequences in said database, a sequence of matching phonetic sequences is automatically generated based on single and optionally multi-syllable phonetic sequences.

56. The system of claim 55, wherein said sequence of matching phonetic sequences is narrowed down through user interaction.

57. The system of claim 55, wherein a sequence of matching ideographic sequences is automatically generated based on matching phonetic sequences to ideographic sequences.

58. The system of claim 57, wherein a sequence of matching ideographic sequences is narrowed down through user interaction.

59. The system of claim 54, wherein a desired phonetic sequence and corresponding ideographic character sequence are specified through a second selection mechanism.

60. The system of claim 38, wherein the user can specify a particular tone for the phonetic syllable.

61. The system of claim 60, wherein one of the plurality of inputs is associated with a special wildcard input that is associated with any or all tones.

62. The system of claim 38, wherein the user can specify an explicit syllable separator.

63. The system of claim 38, wherein once the user enters a sequence of phonetic characters, the user is returned a sequence of phonetic sequences of exact matches and predictions that partially match.

64. The system of claim 63, wherein the sequence is ordered according to the frequency of use based on a linguistic model.

65. The system of claim 64, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of phonetic sequences or ideographic sequences in formal or conversational written text;

frequency of occurrence of phonetic sequences or ideographic when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character sequence entry; and

recency of use or repeated use of phonetic sequences by the user or within an application program.

66. The system of claim 38, wherein once the user has selected a sequence of ideographic characters, the user is presented with a list of sequences of one or more ideographic characters.

67. The system of claim 66, wherein said list of sequences is ordered according to the frequency of use based on a linguistic model.

68. The system of claim 67, where said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of ideographic characters in formal or conversational written text;

frequency of occurrence of ideographic characters when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character entry; and

recency of use or repeated use of ideographic characters by the user or within an application program.

69. An ideographic language text input system incorporated in a user input device, comprising:

a plurality of inputs, each of the plurality of inputs associated with a plurality of characters, an input sequence being generated each time an input is selected by manipulating the user input device, wherein a generated input sequence corresponds to a sequence of inputs that have been selected;

at least one selection input for generating an object output, wherein an input sequence is terminated when the user manipulates the user input device to a selection input;

a memory containing a plurality of objects, wherein each of the plurality of objects is associated with an input sequence;

a display to depict system output to the user; and

a processor coupled to the user input device, memory, and display, said processor comprising;

identifying means for identifying from the plurality of objects in the memory any object associated with each generated input sequence;

output means for displaying on the display the character interpretation of any identified objects associated with each generated input sequence; and

selection means for selecting the desired character for entry into a text entry display location upon detecting the manipulation of the user input device to a selection input.

70. The system of claim 69, wherein said selection means selects a desired character based upon identification of objects having a highest priority based on a linguistic model.

71. The system of claim 69, wherein each time a phrase or ideographic sequence is selected, input sequences for phrases and ideographic sequences that comprise are reprioritized.

72. The system of claim 69, wherein an object is added to a memory if an object does not exist for an input sequence.

73. The system of claim 69, wherein one of the plurality of inputs is associated with a special wildcard input that is associated with any or all tones and delimiter.

74. A system for disambiguating ambiguous input sequences entered by a user and generating textual output in Chinese language, said system comprising:

a user input device having a plurality of input means, each of said input means being associated with a plurality of Latin alphabets, an input sequence being generated each time when an input is selected by said user input device, said generated input sequence having a textual interpretation that is ambiguous due to the plurality of Latin alphabets associated with said inputs;

a memory containing data used to construct a plurality of Pinyin spellings, each of said Pinyin spellings being associated with an input sequence and a frequency of use based on a linguistic model, and each of said Pinyin spellings comprising a sequence of Pinyin syllables corresponding to a phonetic reading to be output to the user, wherein said Pinyin spellings are constructed from data stored in said memory in a tree structure comprised of a plurality of nodes, each of said nodes being associated with an input sequence;

a display to depict system output to the user; and

a processor coupled to said user input device, said memory and said display, said processor constructing a Pinyin spelling from said data in said memory associated with each input sequence and identifying at least one candidate Pinyin spelling with the highest frequency of use based on a linguistic model, and generating an output signal causing said display to display said at least one identified candidate Pinyin spelling associated with each generated input sequence as a textual interpretation of said generated sequence.

75. The system of claim 74, wherein one or more Pinyin spelling objects in said tree structure in memory is associated with one or more Chinese phrases, wherein each Chinese phrase is a textual interpretation of said associated Pinyin spelling object, and wherein each Chinese phrase object is associated with a frequency of use based on a linguistic model.

76. The system of claim 75, wherein said processor constructs at least one identified candidate Chinese phrase for a selected Pinyin spelling and generates an output signal causing said display to display said at least one identified candidate Chinese phrase associated with said selected Pinyin spelling associated with each generated input sequence as a textual interpretation of said generated sequence.

77. The system of claim 76, wherein said at least one identified Chinese phrase has a Pinyin spelling exactly matching said selected Pinyin spelling.

78. The system of claim 76, wherein said at least one identified Chinese phrase has a Pinyin spelling exactly matching all syllables except the last syllable of said selected Pinyin spelling and the last syllable of the Pinyin of said identified Chinese phrase is a completed syllable that can be extended from the last syllable of said selected Pinyin spelling.

79. The system of claim 76, wherein said frequency of use based on a linguistic model associated with each Pinyin spelling object corresponds to a sum of the frequencies of use of all Chinese phrase objects associated with said Pinyin spelling object.

80. The system of claim 79, wherein said Pinyin spelling with the highest frequency of use based on a linguistic model is a default Pinyin spelling selection.

81. The system of claim 75, wherein a Chinese phrase with the highest frequency of use based on a linguistic model is a default Chinese phrase selection.

82. The system of claim 75, wherein at least one or more of said plurality of inputs is an unambiguous navigation input;

and

wherein the user may search for next set of Chinese phrases corresponding to a selected Pinyin spelling as interpretation of an input sequence by additional selections of said navigation inputs, each selection of said unambiguous navigation inputs displaying an alternate list of Chinese phrases corresponding to said selected Pinyin spelling in said memory associated with said generated input sequence.

83. The system of claim 74, wherein at least one or more of said plurality of inputs is an unambiguous navigation input and

wherein the user may select an alternate Pinyin spelling as interpretation of an input sequence by additional selections of said navigation inputs, each selection of said unambiguous navigation inputs selecting a Pinyin spelling object from said identified one or more Pinyin spelling objects in said memory associated with said generated input sequence.

84. The system of claim 74, wherein said user input device comprises an additional input which can be activated to input a tone for a Pinyin syllable.

85. The system of claim 84, wherein one or more Pinyin syllables including tones are associated with the same input with which the corresponding Pinyin syllables are input without tones.

86. The system of claim 85, wherein the tones of each of the Chinese characters are also stored in the memory;

and

wherein only Chinese phrases with characters which have tones matched corresponding input tones are outputted to the user.

87. The system of claim 74, wherein an object is added to a custom database if an object does not exist for an input sequence.

88. The system of claim 87, wherein in absence of matching phonetic sequences in said database, a sequence of matching phonetic sequences is automatically generated based on single and optionally multi-syllable phonetic sequences.

89. The system of claim 88, wherein said sequence of matching phonetic sequences is narrowed down through user interaction.

90. The system of claim 89, wherein a sequence of matching ideographic sequences is automatically generated based on matching phonetic sequences to ideographic sequences.

91. The system of claim 90, wherein a sequence of matching ideographic sequences is narrowed down through user interaction.

92. The system of claim 91, wherein once a selection has been made the matching input sequence, the matching phonetic sequence and the matching ideographic sequence are added to the memory.

93. The system of claim 74, further comprising:

means for changing the associated priority of the matching phonetic sequence, and sequence of ideographic characters once an ideographic character sequence is selected.

94. The system of claim 74, wherein a desired phonetic sequence and corresponding ideographic character sequence are specified through a second selection mechanism.

95. The system of claim 74, wherein one of the plurality of inputs is associated with a special wildcard input that is associated with any or all tones.

96. The system of claim 74, wherein the user can specify an explicit syllable separator.

97. The system of claim 74, wherein once the user enters a sequence of phonetic characters, the user is returned a sequence of phonetic sequences of exact matches and predictions that partially match.

98. The system of claim 97, wherein the sequence is ordered according to the frequency of use based on a linguistic model.

99. The system of claim 98, where said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of phonetic sequences or ideographic sequences in formal or conversational written text;

frequency of occurrence of phonetic sequences or ideographic when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character sequence entry; and

recency of use or repeated use of phonetic sequences by the user or within an application program.

100. The system of claim 74, wherein once the user has selected a sequence of ideographic characters, the user is presented with a list of sequences of one or more ideographic characters.

101. The system of claim 100, wherein said list of sequences is ordered according to the frequency of use based on a linguistic model.

102. The system of claim 101, wherein said linguistic model comprises at least one of:

number of total keystrokes in an ideograph;

radical of an ideograph;

radical and number of strokes of radical;

alphabetically ordered;

frequency of occurrence of ideographic characters in formal or conversational written text;

frequency of occurrence of ideographic characters when following a preceding character or characters;

proper or common grammar of the surrounding sentence;

application context of current character entry; and

recency of use or repeated use of ideographic characters by the user or within an application program.

103. The system of claim 74, wherein the matches between the input sequence and the phonetic sequences are part of confusion sets.

104. The system of claim 103, wherein the user can select which confusion sets are active.

105. The system of claim 104, wherein one of the plurality of inputs is associated with providing alternative phonetic sequence interpretations of the input sequence based on confusion sets or misspellings.

106. The system of claim 103, wherein the system adapts to the common misspellings or confusion sets of the user.