Method and device for ascertaining feature vectors from a signal

• US 7,646,912 B2
• Filed: 02/18/2005
• Issued: 01/12/2010
• Est. Priority Date: 02/19/2004
• Status: Active Grant

• Current Assignee: Infineon Technologies AG
• Original Assignee: Infineon Technologies AG
• Inventors: Hemmert, Werner, Holmberg, Marcus
• Primary Examiner(s): Desire; Gregory M
• Application Number:
  US11/061,108
• Publication Number:
  US 20050232496A1
• Time in Patent Office: 1,789 Days
• Field of Search: 370/208, 370/210, 382/181, 382/224, 382/260, 382/261, 382/263, 382/197, 704/219, 704/226, 704/230, 704/233, 704/234, 704/255, 706/52
• US Class Current: 382/181
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence

Abstract

A signal is used to form intermediate feature vectors which are subjected to high-pass filtering. The high-pass-filtered intermediate feature vectors have a respective prescribed addition feature vector added to them.

Claims

1. A device for the computer-aided ascertainment of feature vectors from a signal for voice recognition, comprising:

a unit for forming intermediate feature vectors from the signal, at least some of the components of the intermediate feature vectors having a power spectrum from part of the digitized signal;

a high-pass filter, which is coupled to the unit for forming intermediate feature vectors, for high-pass filtering the intermediate feature vectors;

a weighting unit, which is coupled to the unit for forming intermediate feature vectors, for multiplying the intermediate feature vectors by a weighting factor; and

a signal addition unit which is set up to add a respective prescribed addition feature vector at least to some of the high-pass-filtered intermediate feature vectors, the addition feature vectors used being the respective intermediate feature vectors multiplied by the weighting factor.

2. The device of claim 1, comprising:

a Wiener filter for providing the digitized signal or for preprocessing the digitized signal or for processing a spectral-transformed digitized signal.

3. A device for computer-aided voice recognition, comprising:

a unit configured for forming intermediate feature vectors from a signal, at least some of the components of the intermediate feature vectors having a power spectrum from part of the signal;

a high-pass filter, which is coupled to the unit for forming intermediate feature vectors, for high-pass filtering the intermediate feature vectors;

a weighting unit, which is coupled to the unit for forming intermediate feature vectors, for multiplying the intermediate feature vectors by a weighting factor;

a signal addition unit which is set up to add a respective prescribed addition feature vector at least to some of the high-pass-filtered intermediate feature vectors, the addition feature vectors used being the respective intermediate feature vectors multiplied by the weighting factor; and

a voice recognition unit for voice recognition using the sum formed as feature vectors.

4. A computer-readable storage medium having a computer program for ascertaining feature vectors from a signal for voice recognition which, when executed by a processor, comprises:

using the signal to form intermediate feature vectors,

subjecting the intermediate feature vectors to high-pass filtering; and

multiplying the intermediate feature vectors by a weighting factor;

wherein at least some of the high-pass-filtered intermediate feature vectors have a respective prescribed addition feature vector added to them, the addition feature vectors used being the respective intermediate feature vectors multiplied by the weighting factor.

5. A computer-readable storage medium having a computer program for computer-aided voice recognition which, when executed by a processor, comprises:

using a signal to form intermediate feature vectors;

subjecting the intermediate feature vectors to high-pass filtering;

multiplying the intermediate feature vectors by a weighting factor;

wherein at least some of the high-pass-filtered intermediate feature vectors have a respective prescribed addition feature vector added to them, the addition feature vectors used being the respective intermediate feature vectors multiplied by the weighting factor; and

using the sum formed as feature vectors to perform voice recognition.

6. A device for computer-aided voice recognition, comprising:

means for forming intermediate feature vectors from a signal, at least some of the components of the intermediate feature vectors having a power spectrum from part of the signal;

a high-pass filter, which is coupled to the means for forming intermediate feature vectors, for high-pass filtering the intermediate feature vectors;

a weighting unit, which is coupled to the means for forming intermediate feature vectors, for multiplying the intermediate feature vectors by a weighting factor;

a signal addition unit which is set up to add a respective prescribed addition feature vector at least to some of the high-pass-filtered intermediate feature vectors, the addition feature vectors used being the respective intermediate feature vectors multiplied by the weighting factor; and

a voice recognition unit for voice recognition using the sum formed as feature vectors.