Skip to main content

Table 3 Statistics for COGNIMUSE database (travel documentaries) annotated with salient events

From: COGNIMUSE: a multimodal video database annotated with saliency, events, semantics and emotion with application to summarization

Layer

LON

RIO

TOK

SYD

GLN

Percentage (%) of salient frames

A

58.7

43.8

60.0

55.1

45.6

V

46.5

48.5

46.6

48.8

40.5

AV

53.9

50.3

54.7

53.7

42.5

AVS

72.7

79.4

80.3

80.4

72.5

Average (pairwise) correlation between annotators

A

0.62

0.25

0.57

0.68

0.52

V

0.50

0.31

0.33

0.51

0.40

AV

0.61

0.33

0.56

0.65

0.43

AVS

0.25

0.08

0.21

0.23

0.28

Chronbach α for three annotators

A

0.83

0.50

0.86

0.80

0.77

V

0.75

0.57

0.76

0.60

0.67

AV

0.83

0.59

0.85

0.79

0.70

AVS

0.51

0.21

0.48

0.45

0.54