Cost-Sensitive Learning for Emotion Robust Speaker Recognition

Figure 3

The frame-level score rank’s probability density functions for target speakers and nontarget speakers over 68 subjects in MASC.