Method and apparatus for speech recognition based on subsyllable spellings

• US 5,222,188 A
• Filed: 08/21/1990
• Issued: 06/22/1993
• Est. Priority Date: 08/21/1990
• Status: Expired due to Fees

• Current Assignee: Emerson & Stern Associates Incorporated San Diego CA
• Original Assignee: Emerson & Stern Associates, Inc.
• Inventors: Hutchins, Sandra E.
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): Doerrler, Michelle
• Application Number:
  US07/570,459
• Time in Patent Office: 1,036 Days
• Field of Search: 381/41-46, 364/513.5
• US Class Current: 704/200
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams

Abstract

In a digital computer, there is provided a method of recognizing speech, comprising the steps of: entering a cohesive speech segment; determining gross acoustic attributes of the entered segment; determining fine acoustic attributes of the entered segment; assigning at least one subsyllable to the entered segment based on the gross and fine acoustic attributes determined; repeating the foregoing steps on successive cohesive speech segments to generate at least one sequence of subsyllables; converting the sequence of subsyllables into a sequence of syllables by finding the sequence of subsyllables in a table in which predetermined subsyllable sequences correspond with respective syllables and syllable sequences; combining the converted sequence of syllables into words; and verifying the conformance of the words to a first predetermined set of grammatical rules. An apparatus implementing the method is also disclosed.

Claims

1. In a digital computer, a method of recognizing speech, comprising the steps of:

entering a cohesive speech segment;

determining gross acoustic attributes of the entered segment;

determining fine acoustic attributes of the entered segment;

assigning at least one subsyllable to the entered segment based on the gross and fine acoustic attributes determined;

repeating the foregoing steps on successive cohesive speech segments to generate at least one sequence of subsyllables;

converting the sequence of subsyllables into a sequence of syllables by finding the sequence of subsyllables in a table in which predetermined subsyllable sequences correspond with respective syllables and syllable sequences;

combining the converted sequence of syllables into words;

verifying the conformance of the words to a stored first predetermined set of grammatical rules; and

reporting a recognition result.

2. A method of recognizing speech as claimed in claim 1, wherein said gross acoustic attributes include silence, fricative, vowel and change interval.

3. A method of recognizing speech as claimed in claim 1, wherein said fine acoustic attributes include duration of a silence, strength or weakness of a fricative, articulatory quality of a vowel, the rising or falling amplitude of a change interval.

4. A method of recognizing speech as claimed in claim 1, wherein a primary and a secondary subsyllable spelling is assigned to the entered segment based on the gross and fine acoustic attributes.

5. A method of recognizing speech as claimed in claim 4, wherein penalty points are assessed to analyzed sequences of speech segments that use a secondary subsyllable to generate a match

6. A method of recognizing speech as claimed in claim 5, wherein the sequence of speech segments with the least number of penalty points is considered the recognized result.

7. A method of recognizing speech as claimed in claim 1, further comprising the step of comparing a string of subsyllables to a predetermined set of subsyllable grammar tables.

8. A method of recognizing speech as claimed in claim 7, wherein said subsyllable grammar tables contain spellings of words or parts of words.

9. A method of recognizing speech as claimed in claim 1, further comprising the step of verifying the conformance of a group of words to a predetermined set of word to phrase grammar rules.

10. An apparatus for speech recognition comprising:

means for entering a cohesive speech segment;

means for determining gross acoustic attributes of the entered segment;

means for determining fine acoustic attributes of the entered segment;

means for assigning at least one subsyllable to the entered segment based on the gross and fine acoustic attributes determined;

means for repeating the foregoing steps on successive cohesive speech segments to generate at least one sequence of subsyllables;

means for converting the sequence of subsyllables into a sequence of syllables by finding the sequence of subsyllables in a table in which determined subsyllable sequences correspond with respective syllables and syllable sequences;

means for combining the converted sequence of syllables into words;

means for verifying the conformance of the words to a stored first predetermined set of grammatical rules; and

means for reporting a recognition result.

11. An apparatus for speech recognition as claimed in claim 10, wherein said gross acoustic attributes include silence, fricative, vowel and change interval.

12. An apparatus for speech recognition as claimed in claim 10, wherein said fine acoustic attributes include duration of a silence, strength or weakness of a fricative, articulatory quality of a vowel, the rising or falling amplitude of a change interval.

13. An apparatus for speech recognition as claimed in claim 10, wherein a primary and a secondary subsyllable spelling is assigned to the entered segment based on the gross and fine acoustic attributes.

14. An apparatus for speech recognition as claimed in claim 13, wherein penalty points are assessed to analyzed sequences of speech segments that use a secondary subsyllable to generate a match.

15. An apparatus for speech recognition as claimed in claim 14, wherein the sequence of speech segments with the least number of penalty points is considered the recognized result.

16. An apparatus for speech recognition as claimed in claim 10, further comprising means for storing subsyllable grammar rules and word to phrase grammar rules in lookup tables.

17. An apparatus for speech recognition as claimed in claim 10, wherein the combining means addresses lookup tables of rules relating to both correct and incorrect grammatical usages of subsyllables and words.

18. An apparatus for speech recognition as claimed in claim 10, further comprising means for comprising a string of subsyllables to said predetermined set of subsyllable grammar tables.

19. An apparatus for speech recognition as claimed in claim 18, wherein said subsyllable grammar table contain spellings of words or parts of words.

20. An apparatus for speech recognition as claimed in claim 10, further comprising a means for verifying the conformance of a group of words to said predetermined set of word to phrase grammar rules.