Table 1 Animal-specific data corpora and distribution.

From: ANIMAL-SPOT enables animal-independent signal detection and classification using deep learning

Dataset

Label type

Target label

Noise label

\(\sum\) labels

smp[#]\(^{1}\)

smp[min.]\(^{2}\)

smp[%]\(^{3}\)

\(\sum\)r[min.]\(^{4}\)

smp[#]\(^{1}\)

smp[min.]\(^{2}\)

smp[%]\(^{3}\)

\(\sum\)r[min.]\(^{4}\)

smp[#]\(^{1}\)

smp[min.]\(^{2}\)

smp[%]\(^{3}\)

\(\sum\)r[min.]\(^{4}\)

Cockatiel*

1271

12.41

40.9

2.46

1840

41.68

59.1

179.10

3111

54.09

100.0

181.56

Sulphur-crested cockatoo

1495

15.26

41.1

3.37

2145

34.99

58.9

46.81

3640

50.25

100.0

50.18

Peach-fronted conure

1174

12.35

55.2

0.22

952

8.88

44.8

4.53

2126

21.23

100.0

4.75

Monk parakeet*

3133

17.20

46.4

0.63

3612

75.25

53.6

12.63

6745

92.45

100.0

13.26

Blue-/golden-winged warbler*

1616

48.99

32.0

5.43

3431

95.10

68.0

14.57

5047

144.09

100.0

20.00

Chinstrap penguin

906

4.86

20.8

0.82

3454

15.40

79.2

3.26

4360

20.26

100.0

4.08

Atlantic cod

382

3.14

30.6

0.19

867

6.00

69.4

20.82

1249

9.14

100.0

21.01

Harbour seal*

2900

55.18

56.4

7.79

2245

58.67

43.6

22.21

5145

113.85

100.0

30.00

Killer whale34*

17,104

649.45

27.8

20.64

44,323

2076.36

72.2

121.49

61,427

2725.81

100.0

142.13

Pygmy pipistrelle*

1570

0.18

31.0

0.10

3490

4.94

69.0

1.11

5060

5.12

100.0

1.21

Chimpanzee*

7079

231.17

57.2

2.89

5305

174.44

42.8

87.11

12,384

405.61

100.0

90.00

BirdVox-Full-Night36

35,402

295.02

50.0

35,402

295.02

50.0

70,804

588

100.0

  1. *Additional noise augmentation training samples:
  2. [1.1] cockatiel—180 (2.87 min.), [1.2] monk parakeet—105 (1.08 min.), [1.3] Blue-/Golden-winged warbler—500 (10.03 min.), [1.4] Harbour seal—2,531 (32.74 min.), [1.5] killer whale—6715 (258.27 min.), [1.6] Pygmy pipistrelle—543 (1.80 min.), [1.7] chimpanzee—1446 (40.34 min.).
  3. 1Samples (smp[#]), 2sample duration in minutes (smp[min.]), 3sample percentage (smp[%]), 4summed target/noise duration of the three unseen test recordings (\(\sum\)r[min.]).
  4. Significant values are in bold.