Method of speech-based navigation in a communications network and of implementing a speech input possibility in private information units

• US 20050102147A1
• Filed: 10/07/2004
• Published: 05/12/2005
• Est. Priority Date: 06/09/1999
• Status: Abandoned Application

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Besling, Stefan, Ullrich, Meinhard, Thelen, Eric
• Application Number:
  US10/960,775
• Publication Number:
  US 20050102147A1
• Time in Patent Office: Days
• US Class Current: 704/270.100
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

The method enables a user of a client (2) to invoke predefined information units in a communications network per speech input. For this purpose, a client (2) downloads a private information unit (27) that enables a speech input from a server (6), a speech recognizer (8) produces a recognition result from an uttered speech input and with the recognition result a link (44-46, 48) to an information unit is determined in a data file (5) to which information unit a word (41-43, 47) is assigned that correlates with the recognition result. Furthermore, with a method of implementing a speech input possibility in private information units (27) for the speech-based navigation in a communications network (4), a registration information unit (19) is downloaded from a server (6) by means of a client (1), by means of which registration information unit (19) user-specific links (46) are assigned to predefined words (41-43) and the assignment (25, 26) with a user identifier (IDn) is transmitted to a data file (5) and the user identifier (IDn) and an address of a speech recognizer (8), which can each be combined with a private information unit (27), are transmitted to the client (1).

Claims

1. A method of speech-based navigation in a communications network (4), comprising the following steps:

a client (2) downloads a private information unit (27) enabling a speech input from a server (6),

a speech recognizer (8) produces a recognition result from an uttered speech input, said speech input recognizer being remotely located from the client (2) and said private information unit subdividing the uttered speech input into feature vectors, wherein said recognition result is performed after the feature vectors are transmitted to the speech recognizer via the server, and

with the recognition result a link (44-46, 48) to an information unit in a data file (5) is determined which is assigned to a word (41-43, 47) that correlates with the recognition result.

2. A method as claimed in claim 1, characterized in that the respective link (44- 46, 48) can be activated via a plurality of speech inputs and/or the respective link (44-46, 48) can be activated via a plurality of speech inputs in different languages.

3. A method as claimed in claim 1, characterized in that the private information unit (27) contains a user identifier (IDn) and the link (44-46, 48) is determined with the recognition result and the user identifier (IDn,) in the data file (5), which link is assigned to a word (41-43, 47) that correlates with the recognition result and is assigned to the user identifier (IDn).

4. A method as claimed in claims 1 and 3, characterized in that the determined link is returned to the client (2) upon its invocation and output.

5. A method as claimed in claims 1 and 3, characterized in that the information unit connected with the determined link is transmitted to the client (2) to be invoked and output.

6. A method as claimed in claim 1, characterized in that the private information unit (27) contains an address of a speech recognizer (8) and the speech recognition is executed on a speech recognition server (3) connected through the communications network (4).

7. A method as claimed in claim 1, characterized in that the speech recognition is executed locally on the client (2).

8.1. 1 A method as claimed in claims 1 and 3, characterized in that the speech recognizer (8) recognizes in addition to predefined words (41-43) also user-independent words (47) to which links (48) are assigned by the service provider and in that, in the case of a speech input correlating with the user-independent words (47) and a recognition result produced therefrom, a user-independent link (48) is determined independently of the user identifier (IDn).

9. A method as claimed in claim 1, characterized in that when the private information unit (27) is invoked, a test is made whether a software module (21) is present on the invoking client (1, 2), which software module is necessary for feature extraction of the speech input and for the transmission to the speech recognizer (8), and this software module, if absent, is downloaded from the server (6).

10. A method as claimed in claim 1, characterized in that at least temporarily an interested party is assigned a right to activate an information unit assigned per link (48) with at least one expression in a natural language that can be recognized by the speech recognizer (8).

13. A method as claimed in claims 1, 8 and 11, characterized in that after the assignment an information unit containing a user identifier (IDn) and an address of a speech recognizer (8) is transmitted to a user for whom no private information unit (27) exists, by means of which information unit the user is enabled to invoke the assigned information units per speech input.

14. A method as claimed in claim 1 or 11, characterized in that the registration information unit (19), the private information unit (27), the speech recognizer (8) and the data file (5) are stored on one (7) or on several servers (3, 5, 6) connected per communications network (4).

15. A software module (21) for implementing the method as claimed in claim 1.

16. A software module as claimed in claim 15, characterized in that it is activated by means of an operating element (24).

17. A computer on which a software module (21) as claimed in claim 15 is executed.

11. A method of implementing a speech input possibility in private information units (27) for speech-based navigation in a communications network (4);

in which a registration information unit (19) is downloaded from the server (6) by means of a client (1), by means of which registration information unit (19) user-specific links (44-46) are assigned to the predefined words (41-43), and the assignment (25, 26) with a user identifier (IDn) is transmitted to a data file (5) and

in which the user identifier (IDn) and an address of a speech recognizer (8) which can each be combined with a private information unit (27) are transmitted to the client (1).

12. A method as claimed in claim 11, characterized in that at least one word (41-43) is combined with a link (44-46) and this assignment (25, 26) is stored in the data file (5) with the individual user identifier (IDn) which each user receives on registration.