Feature extraction using HHT-based locally optimized short-time fractional Fourier transform for speaker recognition | IEEE Conference Publication | IEEE Xplore