Systems and methods for feature extraction

1. A method for feature extraction by an electronic device, comprising:

processing speech using a physiological cochlear model;

analyzing sections of an output of the physiological cochlear model to produce regions;

extracting a place-based analysis vector for each region, wherein extracting a place-based analysis vector comprises averaging a region over time; and

determining one or more features for each place-based analysis vector, wherein determining one or more features comprises determining for a positive part of each place-based analysis vector at least an average and a median, and determining for a negative part of each place-based analysis vector at least an average and a median.

2. The method of claim 1 , further comprising selecting a voiced portion in the output of the cochlear model.

3. The method of claim 1 , further comprising:

determining a foreground quality, a background quality and an overall quality based on foreground distortions and background distortions.

4. The method of claim 3 , further comprising displaying the overall quality.

5. The method of claim 4 , further comprising displaying the foreground quality, the background quality or both.

6. The method of claim 4 , wherein a window over which the overall quality is determined is configurable.

7. The method of claim 1 , wherein the electronic device is a wireless communication device or a network device.

8. The method of claim 1 , wherein extracting the place-based analysis vector is based on an original speech signal, a modified speech signal, or a combination of both.

9. The method of claim 8 , wherein the combination is difference.

10. The method of claim 9 , wherein the difference is a difference of logarithms.

11. The method of claim 1 , further comprising extracting a time-based analysis vector for each sub-region, wherein extracting a time-based analysis vector comprises averaging a region over place.

12. The method of claim 1 , wherein the region over time that is averaged is from a modified speech signal, wherein extracting a place-based analysis vector further comprises determining a logarithm of the averaged region over time for the modified speech signal.

13. The method of claim 12 , wherein extracting a place-based analysis vector further comprises:

averaging another region over time for an original speech signal;

determining a logarithm of the averaged region over time for the original speech signal; and

comparing the logarithm of the averaged region over time for the modified speech signal with the logarithm of the averaged region over time for an original speech signal to produce the place-based analysis vector.

14. The method of claim 1 , wherein determining one or more features further comprises determining at least an average and a median for each placed-based analysis vector.

15. An electronic device for feature extraction, comprising:

physiological cochlear model circuitry configured to process speech;

section analysis circuitry coupled to the physiological cochlear model circuitry, wherein the section analysis circuitry is configured to analyze sections of an output of the physiological cochlear model circuitry to produce regions;

vector extraction circuitry coupled to the section analysis circuitry, wherein the vector extraction circuitry is configured to extract a place-based analysis vector for each region, wherein extracting a place-based analysis vector comprises averaging a region over time; and

feature determination circuitry coupled to the vector extraction circuitry, wherein the feature determination circuitry is configured to determine one or more features for each place-based analysis vector, wherein determining one or more features comprises determining for a positive part of each place-based analysis vector at least an average and a median, and determining for a negative part of each place-based analysis vector at least an average and a median.

16. The electronic device of claim 15 , further comprising portion selection circuitry configured to select a voiced portion in the output of the cochlear model.

17. The electronic device of claim 15 , wherein the electronic device further comprises a foreground quality estimator configured to estimate a foreground quality, a background quality, and an overall quality based on foreground distortions and background distortions.

18. The electronic device of claim 17 , further comprising a display configured to display the overall quality.

19. The electronic device of claim 15 , wherein the electronic device is a wireless communication device or a network device.

20. The electronic device of claim 15 , wherein extracting the place-based analysis vector is based on an original speech signal, a modified speech signal, or a combination of both.

21. The electronic device of claim 15 , wherein the circuitry is integrated in a processor.

22. An apparatus for feature extraction, comprising:

means for processing speech using a physiological cochlear model;

means for analyzing sections of an output of the physiological cochlear model to produce regions;

means for extracting a place-based analysis vector for each region, wherein extracting a place-based analysis vector comprises averaging a region over time; and

means for determining one or more features for each place-based analysis vector, wherein determining one or more features comprises determining for a positive part of each place-based analysis vector at least an average and a median, and determining for a negative part of each place-based analysis vector at least an average and a median.

23. The apparatus of claim 22 , further comprising means for selecting a voiced portion in the output of the cochlear model.

24. A computer-program product for feature extraction, comprising a non-transitory computer-readable medium having instructions thereon, the instructions comprising:

code for causing an electronic device to process speech using a physiological cochlear model;

code for causing the electronic device to analyze sections of an output of the physiological cochlear model to produce regions;

code for causing the electronic device to extract a place-based analysis vector for each region, wherein extracting a place-based analysis vector comprises averaging a region over time; and

code for causing the electronic device to determine one or more features for each place-based analysis vector, wherein determining one or more features comprises determining for a positive part of each place-based analysis vector at least an average and a median, and determining for a negative part of each place-based analysis vector at least an average and a median.

25. The computer-program product of claim 24 , further comprising code for causing the electronic device to select a voiced portion in the output of the cochlear model.