Generation of a reference-model directory for a voice-controlled communications device
First Claim
Patent Images
1. A method for creating a reference model list for a speaker-independent voice recognition system in a voice-controlled communication appliance, the method comprising the steps of:
- generating at least one command word in text form on the communication appliance;
transmitting the at least one command word to a converter station via a communication link initiated by the communication appliance;
creating at least one reference model by the converter station using the at least one command word, with a respective command word having a reference model allocated to it via word/model conversion;
sending the at least one reference model to the terminal via a further communication link; and
storing the at least one reference model received in a reference model list by the communication appliance, wherein the reference model list is accessed by the speaker-independent voice recognition system during operation on the voice-controlled communication appliance.
2 Assignments
0 Petitions
Accused Products
Abstract
For a voice recognition system in a voice-controlled communication appliance, command words from a vocabulary are entered in text form and are transmitted to a separate converter station via a communication link initiated by the communication appliance, wherein the converter station performs word/model conversion on the command words and sends the reference models created in this way via a communication link to the terminal, in which they are stored and are made available to the voice recognition system.
13 Citations
12 Claims
-
1. A method for creating a reference model list for a speaker-independent voice recognition system in a voice-controlled communication appliance, the method comprising the steps of:
-
generating at least one command word in text form on the communication appliance; transmitting the at least one command word to a converter station via a communication link initiated by the communication appliance; creating at least one reference model by the converter station using the at least one command word, with a respective command word having a reference model allocated to it via word/model conversion; sending the at least one reference model to the terminal via a further communication link; and storing the at least one reference model received in a reference model list by the communication appliance, wherein the reference model list is accessed by the speaker-independent voice recognition system during operation on the voice-controlled communication appliance. - View Dependent Claims (2, 3, 4)
-
-
5. An apparatus for creating reference models for speaker-independent voice recognition systems, comprising:
-
a conversion part for performing word/model conversion on command words in text form to produce reference models; and an interface connected to the conversion part that receives command words and passes the command words to the conversion part, wherein the interface transmits reference models created by the conversion part, and wherein the reference models are accessed by speaker-independent voice recognition systems during operation on a voice-controlled communication appliance. - View Dependent Claims (6, 7, 8)
-
-
9. A voice-controlled communication appliance, comprising:
-
a speaker-independent voice recognition system having a memory for a reference model list; an input for entering command words in text form; and a transmission/reception device connected both to the input and to the memory of the reference model list, wherein the transmission/reception device sets up communication links to a converter station which transmits command words and receives reference models that are accessed by the speaker-independent voice recognition system during operation on the voice-controlled communication appliance. - View Dependent Claims (10, 11)
-
-
12. A method for creating a reference model list in a voice-controlled communication appliance, the method comprising the steps of:
-
entering at least one speaker-independent command word in text form on the communication appliance by a user; transmitting the at least one command word to a converter station via a communication link initiated by the communication appliance; receiving at least one reference model transmitted from the converter station using the at least one command word, with a respective command word having a reference model allocated to it via word/model conversion, wherein the at least one reference model is received via a further communication link; and storing the at least one reference model received in a reference model list by the communication appliance, wherein the reference model list is accessed by the speaker-independent voice recognition system during operation on the voice-controlled communication appliance.
-
Specification