Table 1 Average V-Uv (%), Uv-V (%), and overall (%) error rate comparison for each SAD method over the noisy speech dataset in Section 3.1

Oracle

Proposed

0.09

0.25

0.34

LmEMD [36]

0.21

0.42

0.63

NfEMD [39]

0.29

0.49

0.78

FrSAD [40]

0.35

0.56

0.91

MF-SAD [30]

0.41

0.71

1.12

SE-SAD [41]

0.48

0.77

1.25

WnF [39]

0.63

0.85

1.48

15

Proposed

0.13

0.36

0.49

LmEMD [36]

0.33

0.64

0.97

NfEMD [39]

0.46

0.77

1.23

FrSAD [40]

0.56

0.89

1.45

MF-SAD [30]

0.67

1.14

1.81

SE-SAD [41]

0.78

1.24

2.02

WnF [39]

0.97

1.38

2.35

10

Proposed

0.25

0.63

0.88

LmEMD [36]

0.48

0.93

1.41

NfEMD [39]

0.69

1.13

1.82

FrSAD [40]

0.82

1.32

2.14

MF-SAD [30]

0.97

1.68

2.65

SE-SAD [41]

1.45

1.84

2.99

WnF [39]

1.78

2.13

3.91

5

Proposed

0.32

0.87

1.19

LmEMD [36]

0.75

1.47

2.22

NfEMD [39]

1.03

1.75

2.78

FrSAD [40]

1.26

2.02

3.28

MF-SAD [30]

1.48

2.57

4.05

SE-SAD [41]

1.74

2.81

4.55

WnF [39]

2.83

3.19

6.02

0

Proposed

0.41

1.12

1.53

LmEMD [36]

0.96

1.87

2.83

NfEMD [39]

1.33

2.11

3.44

FrSAD [40]

1.66

2.67

4.33

MF-SAD [30]

1.89

2.82

4.71

SE-SAD [41]

2.15

3.26

5.41

WnF [39]

3.11

5.23

8.34

-5

Proposed

0.88

1.94

2.82

LmEMD [36]

1.91

2.82

2.83

NfEMD [39]

2.63

3.71

4.73

FrSAD [40]

3.11

4.17

7.28

MF-SAD [30]

3.83

5.29

9.12

SE-SAD [41]

4.05

5.96

10.01

WnF [39]

5.83

7.11

12.94