Method of speech recognition

• US 6,757,655 B1
• Filed: 08/27/1999
• Issued: 06/29/2004
• Est. Priority Date: 03/09/1999
• Status: Expired due to Term

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Thelen, Eric, Besling, Stefan, Ullrich, Meinhard
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Azad, Abul K.
• Application Number:
  US09/385,634
• Time in Patent Office: 1,768 Days
• Field of Search: 704/270.1, 704/275, 704/271, 704/272, 704/251, 704/231, 709/203, 709/205, 709/226
• US Class Current: 704/270.1
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

In a method in which an information unit (4) enabling a speech input is stored on a server (5) and can be retrieved by a client (1, 2, 3) and in which the client can be coupled to one or more speech recognizers (7, 8, 9) through a communications network (6), the information unit (4) is assigned additional information (12) which is provided for determining a combination of a client (1, 2, 3) for recognizing an uttered speech signal and at least one of the speech recognizers (7, 8, 9), to dynamically assign the speech recognizers (7, 8, 9) in a communications network (6) to the information units (4) and thus ensure an acceptable processing time for the recognition of a speech input with a high recognition quality.

Claims

1. A method in which an information unit is stored on a server and can be retrieved by a client through an input of speech and in which the client can be coupled to at least one speech recognizer having a specialized recognition capability associated with the speech input out of a plurality of speech recognizers having predetermined recognition specialties that have been previously assigned by a provider of the information unit and issued by a server on a communications network with an additional identifying information associated with the speech input, said additional identifying information being used for determining a combination of a client with said at least one speech recognizer for recognizing a signal that has been entered, wherein the at least one speech recognizer having an expected recognizable speech vocabulary associated with said information unit, wherein the additional identifying information includes one of a predetermined theme area, an indicator for specifying a speech recognizer, or parameters for allowing a speech recognizer to adapt to the speech input, so that the speech input by the client is recognized by the communication network and the at least one speech recognizer having a specialized recognition capability associated with the speech input are coupled to maximize speed and quality while downloading the information unit associated with the speech input;

and

combining the additional information and the information unit as a combined information unit and co-transferring combined information unit during downloading so that the information unit is assigned a speech recognizer specially attuned to the downloaded combined information unit;

wherein the client is able to input speech before the contents of the information unit is coupled to the contents of the speech input, and

wherein if more than one client invokes a same speech recognizer, at least one of the clients is assigned another speech recognizer according to alternative addresses of recognizers stored in a tag in the information unit.

2. A method as claimed in claim 1, wherein the additional information contains in addition to an address for assigning the at least one speech recognizer, indications about the processing of the recognition result and/or type of speech recognizer and/or parameters to be transported to the speech recognizer for its adaptation.

3. A method as claimed in claim 1 wherein the additional information is assigned to the information unit by the server.

4. A method as claimed in claim 1, wherein the information unit with the assigned additional information is loaded by the client.

5. A method as claimed in one or more of the claims 1 to 4, wherein an input speech signal is transported for recognition from the client to the address of the at least one speech recognizer, which address is indicated in the additional information.

6. A method as claimed in claim 1, wherein the additional information contains an address of a distributor to which the input speech signal is conveyed, and is assigned to at least one speech recognizer of the plurality of speech recognizers.

7. A method as claimed in claim 1 wherein the information unit is realized as an HTML page and the additional information as an HTML tag.

8. A method as claimed in one of the claims 1 to 7, wherein there is multiple access to one HTML page by different clients, each client is assigned various address of the plurality of speech recognizers in the HTML tag.

9. A method as claimed in one of the claims 1 or 7, wherein the client has additional software for a feature extraction of a speech signal applied to the client which software is started when the HTML page include the HTML tag is downloaded.

10. A method as claimed in one or more of the claims 1 to 9, wherein a feature stream generated during the feature extraction with respect to the speech signal is applied to the at least one speech recognizer defined in the HTML

at which speech recognizer executes the speech recognition and returns the recognition result to the client in dependence on indications contained in the HTML tag.

11. A method as claimed in claim 10, wherein the recognition results is returned to the server in dependence on indications contained in the HTML tag.

12. A server on which an information unit is stored and can be retrieved by a client, while the client can be coupled to one or more speech recognizers having an expected recognizable speech vocabulary associated with said information unit, wherein additional identifying information selected by the provider of the information unit is assigned to the information unit by the server for determining a combination between the client and said one or more speech recognizers for recognizing an input speech signal associated with said information unit, wherein the additional information includes information comprising one of a predetermined theme area, an indicator for specifying a speech recognizer, or parameters for allowing a speech recognizer to adapt to the speech input;

wherein the additional information and the information unit are combined by the server into a combined information unit and co-transferred during downloading so that the information unit is assigned a speech recognizer specially attuned to the downloaded combined information unit;

wherein the client is able to input speech before the contents of the information unit is coupled to the contents of the speech input, and

wherein if more than one client invokes a same speech recognizer, at least one of the clients is assigned another speech recognizer according to alternative addresses of recognizers stored in a tag in the information unit.

13. A client which can be coupled to one or more speech recognizers through a communications network and which is provided for retrieving an information unit stored on a server and additional information is combined with the information unit for determining a combination between the client and said one or more speech recognizers having an expected recognizable speech vocabulary associated with said information unit and for recognizing an input speech signal, wherein the additional information includes information comprising one of a predetermined theme area, an indicator for specifying a speech recognizer, or parameters for allowing a speech recognizer to adapt to the speech input, and wherein

the client receives the additional information and the information unit as a combined information unit that has been co-transferred during downloading so that the information unit is assigned a speech recognizer specially attuned to the downloaded combined information unit;

wherein the client is able to input speech before the contents of the information unit is coupled to the contents of the speech input, and

wherein if more than one client invokes a same speech recognizer, at least one of the clients is assigned another speech recognizer according to alternative addresses of recognizers stored in a tag in the information unit.