公开(公告)号：MXPA06015240A

公开(公告)日：2007-03-26

申请号：MXPA06015240

申请日：2005-06-06

Applicant: MOTOROLA INC

Inventor： KUSHNER WILLIAM M ,  HARTON SARA M ,  JASIUK MARK A

IPC: G10L21/02 ,  H04B15/00

Abstract: Un metodo para detectar y atenuar ruido de inhalacion en un sistema de comunicacion acoplado a un sistema de suministro de aire presurizado, el metodo incluye los pasos de: generar un modelo de ruido de inhalacion (912, 1012) con base en el ruido de inhalacion; recibir una sefial de entrada (802) que incluya ruido de inhalacion; comparar (810) la sefial de entrada con el modelo de ruido para obtener una medida de similitud; determinar (854) un factor de ganancia con base en la medida de similitud; y modificar (852) la senal de entrada con base en el factor de ganancia, en donde el ruido de inhalacion en la senal de entrada es atenuado con base en el factor de ganancia.

公开(公告)号：GB2333877A

公开(公告)日：1999-08-04

申请号：GB9900679

申请日：1999-01-14

Applicant: MOTOROLA INC

Inventor： SRENGER EDWARD ,  MEUNIER JEFFREY A ,  KUSHNER WILLIAM M

IPC: G10L15/06 ,  G10L5/06

Abstract: A standard filler, or garbage model, is calculated for the detection of out of vocabulary utterances. Gobal statistical parameters are calculated (116) based upon the statistical parameters for new training data and a garbage model (118) is updated based upon the global statistical parameters. This is carried out on-line while the user is enrolling the vocabulary. The garbage model is preferably an average speaker model, representative of all the speech data enrolled by the user to date. The garbage model is preferably obtained as a by-product of the vocabulary enrollment procedure and is similar in it characteristics and topology to all the other regular vocabulary HMMs.

公开(公告)号：CA2572152A1

公开(公告)日：2006-01-19

申请号：CA2572152

申请日：2005-06-09

Applicant: MOTOROLA INC

Inventor： KUSHNER WILLIAM M ,  JASIUK MARK A ,  HARTON SARA M

IPC: G10L15/00 ,  G10L15/06 ,  G10L21/00

Abstract: A method for characterizing inhalation noise within a pressurized air delivery system, the method including the steps of: generating an inhalation noise model ( 912, 1012 ) based on inhalation noise; receiving an input signal ( 802 ) that includes inhalation noise comprising at least one inhalation noise burst; comparing ( 810 ) the input signal to the noise model to obtain a similarity measure; comparing the similarity measure to at least one threshold ( 832, 834 ) to detect the at least one inhalation noise burst; and characterizing ( 1354, 1356 ) the at least one detected inhalation noise burst.

公开(公告)号：CA2571768A1

公开(公告)日：2006-01-19

申请号：CA2571768

申请日：2005-06-06

Applicant: MOTOROLA INC

Inventor： HARTON SARA M ,  JASIUK MARK A ,  KUSHNER WILLIAM M

IPC: G10L21/02 ,  H04B15/00

Abstract: A method for detecting and attenuating inhalation noise in a communication system coupled to a pressurized air delivery system, the method including the steps of: generating an inhalation noise model ( 912, 1012 ) based on inhalation noise; receiving an input signal ( 802 ) that includes inhalation noise; comparing ( 810 ) the input signal to the noise model to obtain a similarity measure; determining ( 854 ) a gain factor based on the similarity measure; and modifying ( 852 ) the input signal based on the gain factor, wherein the inhalation noise in the input signal is attenuated based on the gain factor.

公开(公告)号：AU2005262623A1

公开(公告)日：2006-01-19

申请号：AU2005262623

申请日：2005-06-06

Applicant: MOTOROLA INC

Inventor： HARTON SARA M ,  JASIUK MARK A ,  KUSHNER WILLIAM M

IPC: G10L15/20 ,  G10L21/02

Abstract: A method for equalizing a speech signal generated within a pressurized air delivery system, the method including the steps of: generating an inhalation noise model ( 1152 ) based on inhalation noise; receiving an input signal ( 802 ) that includes a speech signal; and equalizing the speech signal ( 1156 ) based on the noise model.

公开(公告)号：AU2005262586A1

公开(公告)日：2006-01-19

申请号：AU2005262586

申请日：2005-06-09

Applicant: MOTOROLA INC

Inventor： KUSHNER WILLIAM M ,  JASIUK MARK A ,  HARTON SARA M

IPC: G10L15/00

Abstract: A method for characterizing inhalation noise within a pressurized air delivery system, the method including the steps of: generating an inhalation noise model ( 912, 1012 ) based on inhalation noise; receiving an input signal ( 802 ) that includes inhalation noise comprising at least one inhalation noise burst; comparing ( 810 ) the input signal to the noise model to obtain a similarity measure; comparing the similarity measure to at least one threshold ( 832, 834 ) to detect the at least one inhalation noise burst; and characterizing ( 1354, 1356 ) the at least one detected inhalation noise burst.

公开(公告)号：CA2572715C

公开(公告)日：2012-04-03

申请号：CA2572715

申请日：2005-06-06

Applicant: MOTOROLA INC

Inventor： KUSHNER WILLIAM M ,  HARTON SARA M ,  JASIUK MARK A

IPC: G10L21/02 ,  G10L15/20

Abstract: A method for equalizing a speech signal generated within a pressurized air delivery system, the method including the steps of: generating an inhalation noise model (1152) based on inhalation noise; receiving an input signal (802) that includes a speech signal; and equalizing the speech signal (1156) based on the noise model.

公开(公告)号：DE602005008974D1

公开(公告)日：2008-09-25

申请号：DE602005008974

申请日：2005-06-06

Applicant: MOTOROLA INC

Inventor： KUSHNER WILLIAM M ,  HARTON SARA M ,  JASIUK MARK A

IPC: G10L21/02 ,  B63C11/26 ,  G10L11/02 ,  G10L15/06 ,  G10L15/20

Abstract: A method for equalizing a speech signal generated within a pressurized air delivery system, the method including the steps of: generating an inhalation noise model ( 1152 ) based on inhalation noise; receiving an input signal ( 802 ) that includes a speech signal; and equalizing the speech signal ( 1156 ) based on the noise model.

公开(公告)号：AU681058B2

公开(公告)日：1997-08-14

申请号：AU5353196

申请日：1996-01-29

Applicant: MOTOROLA INC

Inventor： HARTMAN MATTHEW A ,  KUSHNER WILLIAM M ,  SRENGER EDWARD

IPC: G10L15/14 ,  G10L9/00

Abstract: In a statistical based speech recognition system, one of the key issues is the selection of the Hidden Markov Model that best matches a given sequence of feature observations. The problem is usually addressed by the calculation of the maximum likelihood, ML, state sequence by means of a Viterbi or other decoder. Noise or inadequate training can produce a ML sequence associated with a Hidden Markov Model other than the correct model. The method of the present invention provides improved robustness by combining the standard ML state sequence score (416) with an additional path score (418) derived from the dynamics of the ML score as a function of time. These two scores, when combined, form a hybrid metric (420) that, when used with the decoder, optimizes selection of the correct Hidden Markov Model (422).

公开(公告)号：AU5353196A

公开(公告)日：1996-10-16

申请号：AU5353196

申请日：1996-01-29

Applicant: MOTOROLA INC

Inventor： HARTMAN MATTHEW A ,  KUSHNER WILLIAM M ,  SRENGER EDWARD

IPC: G10L15/14 ,  G10L9/00

Abstract: In a statistical based speech recognition system, one of the key issues is the selection of the Hidden Markov Model that best matches a given sequence of feature observations. The problem is usually addressed by the calculation of the maximum likelihood, ML, state sequence by means of a Viterbi or other decoder. Noise or inadequate training can produce a ML sequence associated with a Hidden Markov Model other than the correct model. The method of the present invention provides improved robustness by combining the standard ML state sequence score (416) with an additional path score (418) derived from the dynamics of the ML score as a function of time. These two scores, when combined, form a hybrid metric (420) that, when used with the decoder, optimizes selection of the correct Hidden Markov Model (422).