Speaker Recognition Using Wavelet Packet Entropy, I-Vector, and Cosine Distance Scoring

Figure 2

The wavelet packet transform at 2 levels.