Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing

Table 2 Miss and FA of LCM-Ivec-Hybrid system for RT09

From: Latent class model with application to speaker diarization

 

Miss[%]

FA[%]

EDI_20071128-1000

3.64

4.81

EDI_20071128-1500

8.36

6.68

IDI_20090128-1600

4.09

1.32

IDI_20090129-1000

5.91

7.78

NIST_20080201-1405

20.01

2.54

NIST_20080227-1501

8.86

1.26

NIST_20080307-0955

5.35

2.49

Average

8.03

3.84

  1. Miss and FA are caused by VAD error and overlapping speech. They are very similar for all the three proposed systems, as the same VAD method is used
Back to article page