Local spectral variability features for speaker verification

This is an implementation of local spectral variability features.

This package contains a new feature for speaker recognition. The feature represents local spectral variability of a speech signal. The variability is computed from spectral features such as MFCCs.
If you find this code package useful and would like to cite it in your work, please cite the original paper that describes the method:

Md Sahidullah, Tomi Kinnunen, Local spectral variability features for speaker verification, Digital Signal Processing, Volume 50, March 2016, Pages 1-11, ISSN 1051-2004, http://dx.doi.org/10.1016/j.dsp.2015.10.011.