Speaker Recognition Using Wavelet Cepstral Coefficient, I-Vector, and Cosine Distance Scoring and Its Application for Forensics

<div>(a) Two speech signals that have different sound volume before normalization. (b) Two speech signals that have different sound volume after normalization.</div>

Journal of Electrical and Computer Engineering

fig3

Figure 3

Figure 3: Speaker Recognition Using Wavelet Cepstral Coefficient, I-Vector, and Cosine Distance Scoring and Its Application for Forensics 