User model-improvement-data-driven selection and update of user-oriented recognition model of a given type for word recognition at network server

• US 6,363,348 B1
• Filed: 10/16/1998
• Issued: 03/26/2002
• Est. Priority Date: 10/20/1997
• Status: Expired due to Term

• Current Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Original Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Inventors: Thelen, Eric, Besling, Stefan
• Primary Examiner(s): SMITS, TALIVALDIS IVARS
• Application Number:
  US09/174,167
• Time in Patent Office: 1,257 Days
• Field of Search: 704/251, 704/275, 704/270.1
• US Class Current: 704/270.1
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 15/34 Adaptation of a single recogniser for parallel processing, e.g. by use of multiple processors or cloud computing

Abstract

A distributed pattern recognition system includes at least one user station and a server station. The server station and the user station are connected via a network, such as Internet. The server station includes different recognition models of a same type. As part of a recognition enrolment, the user station transfers model improvement data associated with a user of the user station to the server station. The server station selects a recognition model from the different recognition models of a same type in dependence on the model improvement data. For each recognition session, the user station transfers an input pattern representative of time sequential input generated by the user to the server station. The server station retrieves a recognition model selected for the user and provides the retrieved recognition model to a recognition unit for recognising the input pattern using the recognition models.

Claims

1. A method for recognizing an input pattern stored in a user station using a recognition unit of a server station;

the server station and the user station being connected via a network;

the recognition unit being operative to recognize the input pattern using a model collection of at least one recognition model;

the method comprising;

performing an initial recognition enrolment step, comprising transferring model improvement data associated with a user of the user station from the user station to the recognition unit; and

associating the user of the user station with a user identifier; and

for a recognition session between the user station and the server station, transferring a user identifier associated with a user of the user station and an input pattern representative of time sequential input generated by the user from the user station to the server station; and

using the recognition unit to recognize the input pattern by incorporating at least one recognition model in the model collection which reflects the model improvement data associated with the user;

characterized;

in that the server comprises a plurality of different user-oriented recognition model sets, each set of a same type;

in that the recognition enrolment step comprises selecting a user-oriented recognition model from a set of a same type in dependence on the model improvement data associated with the user; and

storing an indication of the selected user-oriented recognition model in association with the user identifier; and

in that the step of recognizing the input pattern comprises retrieving a user-oriented recognition model associated with the user identifier transferred to the server station and incorporating the retrieved user-oriented recognition model in the model collection.

2. A method as claimed in claim 1, characterized:

in that the plurality of user-oriented recognition models of a same type is formed by a basic recognition model and a plurality of adaptation profiles;

in that the step of selecting a user-oriented recognition model in dependence on the model improvement data associated with the user comprises selecting at least one of the adaptation profiles in dependence on the model improvement data; and

in that the recognition enrolment step comprises storing an indication of the selected adaptation profile in association with the user identifier; and

in that the step of retrieving a user-oriented recognition model associated with the user identifier comprises retrieving an adaptation profile associated with the user identifier and adapting the basic recognition model under control of the adaptation profile.

3. A method as claimed in claim 1, characterized in that the input pattern comprises speech representative data;

in that the model improvement data comprises acoustic training data and that selecting a recognition model from the plurality of different user-oriented recognition models comprises, based on the acoustic training data associated with the user, selecting an acoustic model from a plurality of different acoustic models or selecting an acoustic model adaptation profile and using the selected acoustic model adaptation profile to adapt a basic acoustic model.

4. A method as claimed in claim 3, characterized in that the acoustic model adaptation profile comprises a matrix for transforming an acoustic references space;

or a set of acoustic references to be combined with acoustic references used by the basic acoustic model.

5. A method as claimed in claim 1, characterized in that the model improvement data comprises language model training data and that selecting a user-oriented recognition model from the plurality of different user-oriented recognition models comprises, based on the language model training data associated with the user, selecting a language model from a plurality of different language models or selecting a language model adaptation profile and using the selected language model adaptation profile to adapt a basic language model.

6. A method as claimed in claim 5, characterized the language model training data comprises at least one context identifier;

and in that the method comprises the step of in the server station selecting a language model or a language model adaptation profile corresponding to the context identifier.

7. A method as claimed in claim 1, characterized in that the model improvement data comprises vocabulary training data and that selecting a recognition model from the plurality of different user-oriented recognition models comprises, based on the vocabulary training data associated with the user, selecting a vocabulary from a plurality of different vocabularies or selecting a vocabulary adaptation profile and using the selected vocabulary adaptation profile to adapt a basic vocabulary.

8. A method as claimed in claim 7, characterized in that the vocabulary training data comprises at least one context identifier;

and in that the method comprises the step of in the server station selecting a vocabulary or a vocabulary adaptation profile corresponding to the context identifier.

9. A method as claimed in claim 6, characterized in that the context identifier comprises a keyword.

10. A method as claimed in claim 9, characterized in that the context identifier comprises or indicates a sequence of words, and in that the method comprises extracting at least one keyword from the sequence of words and performing the selection based on the extracted keyword(s).

11. A pattern recognition system comprising at least one user station storing an input pattern and a server station comprising a recognition unit;

the recognition unit being operative to recognize the input pattern using a model collection of at least one recognition model;

the server station being connected to the user station via a network;

the user station comprising means for;

initially transferring model improvement data associated with a user of the user station and a user identifier associated with the user to the server station; and

for each recognition session between the user station and the server station transferring a user identifier associated with a user of the user station and an input pattern representative of time sequential input generated by the user to the server station; and

the server station comprising means for, for each recognition session between the user station and the server station,

incorporating at least one recognition model in the model collection which reflects the model improvement data associated with a user from which the input pattern originated; and

using the speech recognition unit to recognize the input pattern received from the user station;

characterized in that the server station comprises;

a plurality of different user-oriented recognition model sets, each set of a same type;

means for selecting a user-oriented recognition model from a a set of a same type in dependence on the model improvement data associated with the user identifier; and

means for retrieving a user-oriented recognition model associated with the user identifier transferred to the server station and for incorporating the retrieved user-oriented recognition model in the model collection.