Motorola Corporation

Yuan-Jun Wei

Changxue Ma

Chen Liu

Steven J Nowlan

Yan Ming Cheng

James R Talley

Ted Mazurkiewicz

Lynette Melnar

Jim Talley

Acoustic phones (preferably drawnfrom a plurality of spoken languages) are providedA hierarchically-organized polyphone network () organizes views of these phones of varying resolution and phone categorization as a function, at least in part, of phonetic similarity () and at least one language-independent phonological factor (). In a preferred approach, a unique transcription system serves to represent the phones using only standard, printable ASCII characters, none of which comprises a special character (such as those characters that have a command significance for common script interpreters such as the UNIX command line).

Polyphone network method and apparatus

An electronic device () for speech dialog includes functions that receive () a speech phrase that comprises a request phrase that includes an instantiated variable (), generate () pitch and voicing characteristics () of the instantiated variable, and performs speech recognition () of the instantiated variable to determine a most likely set of acoustic states (). The electronic device may generate () a synthesized value of the instantiated variable using the most likely set of acoustic states and the pitch and voicing characteristics of the instantiated variable. The electronic device may use a table of previously entered values of variables that have been determined to be unique, and in which the values are associated with a most likely set of acoustic states and the pitch and voicing characteristics determined at the receipt of each value to disambiguate () a newly received instantiated variable.

Speech dialog method and system

The invention provides a Hidden Markov Model () based automated speech recognition system () that dynamically adapts to changing background noise by detecting long pauses in speech, and for each pause processing background noise during the pause to extract a feature vector that characterizes the background noise, identifying a Gaussian mixture component of noise states that most closely matches the extracted feature vector, and updating the mean of the identified Gaussian mixture component so that it more closely matches the extracted feature vector, and consequently more closely matches the current noise environment. Alternatively, the process is also applied to refine the Gaussian mixtures associated with other emitting states of the Hidden Markov Model.

Speech recognition by dynamical noise model adaptation

Growing community of inventors

Hoffman Estates, IL, United States of America

Yuan-Jun Wei