Method, module, device and server for voice recognition

• US 7,983,911 B2
• Filed: 02/12/2002
• Issued: 07/19/2011
• Est. Priority Date: 02/13/2001
• Status: Expired due to Fees

• Current Assignee: InterDigital Patent Holdings, Inc. (InterDigital, Inc.)
• Original Assignee: Thomson Licensing (Vantiva SA)
• Inventors: Soufflet, Frédéric, Tazine, Nour-Eddine
• Primary Examiner(s): Wozniak; James S
• Assistant Examiner(s): He; Jialong
• Application Number:
  US10/467,586
• Publication Number:
  US 20050102142A1
• Time in Patent Office: 3,444 Days
• Field of Search: 704/235, 704/260, 704/257, 704/270.1, 704/231, 704/236, 704/251
• US Class Current: 704/251
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

The invention relates to a speech recognition process implemented in at least one terminal (114), the speech recognition process using a language model (311), comprising the following steps:

• • detection (502) of at least one unrecognized expression in one of the terminals;
  • recording (503) in the terminal of data representative of the unrecognized expression (309);
  • transmission (603) by the terminal of the recorded data to a remote server (116);
  • analysis, (803) at the level of the remote server, of the data and generation (805) of information for correcting the said language model taking account of at least one part of the unrecognized expression; and
  • transmission (806) from the server to at least one terminal (114, 117, 118) of the correcting information, so as to allow future recognition of at least certain of the unrecognized expressions.

The invention also relates to corresponding modules, devices (102) and a remote server (116).

Claims

1. Speech recognition process implemented in a terminal and a remote server, said terminal comprising a language model, comprising the steps of

detection of an expression unrecognized by said language model;

recording in said terminal of data representative of said unrecognized expression;

repetition of the step of detection of an unrecognized expression and the step of recording of representative data, the repetition allowing recording of data representative of a plurality of unrecognized expressions;

transmission to said remote server of recorded data representative of the plurality of unrecognized expressions and at least one of context data and information related to a speaker;

analysis, by said remote server, of said data representative of said plurality of unrecognized expressions and at least one of said context data and said information related to a speaker, and generating information recognizing said unrecognized expressions, said generated information including correcting information to partially or completely modify said language model by adding, deleting or replacing elements therein, an unrecognized expression having no meaning being ignored for the updating of the language model and being discarded from the correcting information for updating said language model;

transmitting, by said server, said correcting information to said terminal; and

updating, by said terminal, said language model.

2. Speech recognition process according to claim 1, said data representative of said unrecognized expressions comprising a compressed speech recording representative of parameters descriptive of an acoustic signal.

3. Speech recognition process according to claim 1, comprising a step of transmission, by said terminal, of said recorded data representative of unrecognized expressions to said remote server;

during said step of transmission by said terminal, the speech recognition process furthermore transmitting to said remote server an item of information forming part of a group comprising;

information of context of use of said speech recognition process when said expression has not been recognized; and

information relating to a speaker who has uttered said unrecognized expression.

4. Speech recognition process according to claim 1, in which the process implements one of an encryption and a scrambling of said recorded data.

5. Speech recognition system using a language model in a terminal and a remote server, comprising:

an analyser detecting expressions unrecognized by said language model;

a recorder for recording data representative of said unrecognized expressions;

a transmitter comprising means for transmitting to said remote server contents of said recorder, including recorded data representative of said unrecognized expressions, and at least one of context data and information related to a speaker;

said remote server comprising means for making comprehensible at least some of said expressions unrecognized by said language model, generating information and updating said language model, said generated information including correcting information to partially or completely modify said language model by adding, deleting or replacing elements therein, an unrecognized expression having no meaning being ignored for the updating of the language model and being discarded from the information for updating;

said remote server comprising means for transmitting said correcting information to said language model for updating said language model so as to allow said language model to recognize said unrecognized expressions.