Cost-Sensitive Learning for Emotion Robust Speaker Recognition

Figure 1

Example of segment boundaries estimation for the phrase “shi de.” The vertical bars represent the segment boundaries from the critical points of pitch contours.