公开(公告)号：DE69905030D1

公开(公告)日：2003-02-27

申请号：DE69905030

申请日：1999-04-21

Applicant: IBM ,  IBM DEUTSCHLAND INFORMATIONSSY

Inventor： FISCHER VOLKER ,  GAO YUQING ,  PICHENY A ,  KUNZMANN SIEGFRIED

IPC: G10L15/07 ,  G10L15/06

公开(公告)号：AU2003293646A1

公开(公告)日：2004-07-14

申请号：AU2003293646

申请日：2003-10-31

Applicant: IBM

Inventor： FISCHER VOLKER ,  KUNZMANN SIEGFRIED

IPC: G10L15/20 ,  G10L15/22 ,  G10L15/26

Abstract: A method and respective system for operating a speech recognition system, in which a plurality of recognizer programs are accessible to be activated for speech recognition, and are combined on a per need basis in order to efficiently improve the results of speech recognition done by a single recognizer. In order to adapt such system to the dynamically changing acoustic conditions of various operating environments and to the particular requirements of running in embedded systems having only a limited computing power available, it is proposed to a) collect selection base data characterizing speech recognition boundary conditions, e.g. the speaker person and the environmental noise, etc., with sensor means, and b) using program-controlled arbiter means for evaluating the collected data, e.g., a decision engine including software mechanism and a physical sensor, to select the best suited recognizer or a combination thereof out of the plurality of available recognizers.

公开(公告)号：DE69905030T2

公开(公告)日：2003-11-27

申请号：DE69905030

申请日：1999-04-21

Applicant: IBM

Inventor： FISCHER VOLKER ,  GAO YUQING ,  PICHENY A ,  KUNZMANN SIEGFRIED

IPC: G10L15/07 ,  G10L15/06

公开(公告)号：AT231642T

公开(公告)日：2003-02-15

申请号：AT99924814

申请日：1999-04-21

Applicant: IBM

Inventor： FISCHER VOLKER ,  GAO YUQING ,  PICHENY MICHAEL A ,  KUNZMANN SIEGFRIED

IPC: G10L15/07 ,  G10L15/06

公开(公告)号：CA2507999C

公开(公告)日：2013-09-03

申请号：CA2507999

申请日：2003-10-31

Applicant: IBM

Inventor： FISCHER VOLKER ,  KUNZMANN SIEGFRIED

IPC: G10L15/20 ,  G10L15/22 ,  G10L15/26

Abstract: The present invention relates to a method and respective system for operating a speech recognition system, in which a plurality of recognizer programs are accessible to be activated for speech recognition, and are combined on a per need basis in order to efficiently improve the results of speech recognition done by a single recognizer. To adapt to dynamically changing acoustic conditions of various operating environments and to embedded systems having only a limited computing power available, it is proposed to a) collect (210,220,230,240) selection base data characterizing speech recognition boundary conditions, e.g. the speaker person and the environmental noise, etc., with sensor means, b) using (260) program-controlled arbiter means for evaluating the collected data, e.g., a decision engine including software mechanism and a physical sensor, to select (290) the best suited recognizer or a combination thereof out of the plurality of available recognizers.

公开(公告)号：DE602006002431D1

公开(公告)日：2008-10-09

申请号：DE602006002431

申请日：2006-05-05

Applicant: IBM

Inventor： KUNZMANN SIEGFRIED ,  FISCHER VOLKER

IPC: G10L13/06 ,  G06F17/30 ,  G10L15/22

公开(公告)号：AT406648T

公开(公告)日：2008-09-15

申请号：AT06113577

申请日：2006-05-05

Applicant: IBM

Inventor： KUNZMANN SIEGFRIED ,  FISCHER VOLKER

IPC: G10L13/06 ,  G06F17/30 ,  G10L15/22

公开(公告)号：CA2507999A1

公开(公告)日：2004-07-08

申请号：CA2507999

申请日：2003-10-31

Applicant: IBM

Inventor： FISCHER VOLKER ,  KUNZMANN SIEGFRIED

IPC: G10L15/20 ,  G10L15/22 ,  G10L15/26

Abstract: The present invention relates to a method and respective system for operatin g a speech recognition system, in which a plurality of recognizer programs are accessible to be activated for speech recognition, and are combined on a per need basis in order to efficiently improve the results of speech recognition done by a single recognizer. To adapt to dynamically changing acoustic conditions of various operating environments and to embedded systems having only a limited computing power available, it is proposed to a) collect (210,220,230,240) selection base data characterizing speech recognition boundary conditions, e.g. the speaker person and the environmental noise, etc., with sensor means, b) using (260) program-controlled arbiter means for evaluating the collected data, e.g., a decision engine including software mechanism and a physical sensor, to select (290) the best suited recognizer or a combination thereof out of the plurality of available recognizers.