Cost-Sensitive Learning for Emotion Robust Speaker Recognition

Figure 2

The probability distribution of PEM for the male (a) and female speakers (b) under the five emotion states.

(a) Male speaker

(b) Female speaker