Skip to main content

Table 4 Accuracy (%) of seal calls recognition by the GRNN model

From: Seal call recognition based on general regression neural network using Mel-frequency cepstrum coefficient features

SNR/dB

MFCC

LOFAR spectrum

A

B

C

Total

A

B

C

Total

20

99.12

97.58

94.01

96.90

95.25

92.32

91.05

92.87

15

97.83

96.02

93.91

95.92

91.71

91.35

90.69

91.25

10

93.55

95.53

91.24

93.44

92.96

89.23

89.28

90.49

5

91.96

94.67

89.37

92.00

90.17

85.82

83.21

86.40

0

88.01

90.65

86.47

88.38

84.42

81.71

79.81

81.98