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

Table 5 Analysis on human annotated dataset

From: Language agnostic missing subtitle detection

Segment level metric

VAD

VAD + AC

ASR

Coverage between predicted speech to reference missing speech sections

72.15

74.91

60.53

Coverage between predicted speech to reference non-speech sections

45.22

35.7

39.3

Back to article page