Acronyms ..................................................... XIII
1 Introduction ............................................... 1
1.1 Why should we care about biodiversity? ..................... 1
1.2 Role of computational bioacoustics ......................... 3
1.3 The benefits of computational bioacoustics ................. 5
1.4 Book organization .......................................... 7
1.5 How to use this book ...................................... 10
1.6 The home page for this book ............................... 12
2 Why computational bioacoustics? ........................... 13
Introduction .............................................. 13
2.1 Ancient roots ............................................. 15
2.2 Contemporary bioacoustics ................................. 17
2.3 Computational bioacoustics ................................ 22
2.4 Relations between bioacoustics and computational
bioacoustics .............................................. 25
2.5 Research topics and application areas ..................... 28
3 Definition of tasks ....................................... 32
Introduction .............................................. 32
3.1 One-species detection ..................................... 37
3.2 Species identification .................................... 39
3.3 Multi-label species identification ........................ 41
3.4 One-category recognition .................................. 45
3.5 One-species recognition ................................... 47
3.6 Multispecies diarization .................................. 49
3.7 Localization and tracking of individuals .................. 52
3.8 Sound event-type recognition .............................. 54
3.9 Abundance assessment ...................................... 57
3.10 Clustering of sound events ................................ 58
3.11 Biodiversity assessment indices ........................... 60
3.12 Practical difficulties related to real-world operation ....
4 Acoustic libraries ........................................ 64
Introduction .............................................. 64
4.1 The training library ...................................... 68
4.1.1 TL organization .................................... 68
4.1.2 Creation of training libraries ..................... 71
4.1.3 Manual creation of TLs ............................. 72
4.1.4 TLs based on clean recordings ...................... 75
4.1.5 Challenges related to TLs' creation ................ 76
4.1.6 Computational bioacoustics in support of TL
creation ........................................... 77
4.1.7 What if large TLs are not feasible? ................ 77
4.2 The acoustic background library ........................... 77
4.2.1 What if large BL is not available? ................. 79
4.2.2 Computational bioacoustics in support of BL
creation ........................................... 80
4.3 The validation library .................................... 82
4.3.1 Computational bioacoustics in support of VL
creation ........................................... 85
4.4 The evaluation library .................................... 86
4.4.1 Computational bioacoustics in support of EL
creation ........................................... 87
Concluding remarks ........................................ 88
5 One-dimensional audio parameterization .................... 89
Introduction .............................................. 89
5.1 Audio pre-processing ...................................... 90
5.1.1 Variable-length segmentation ....................... 92
5.1.2 Uniform-length segmentation ........................ 93
5.2 Audio parameterization .................................... 93
5.3 Post-processing of audio features ......................... 96
5.3.1 Statistical standardization ........................ 98
5.3.2 Temporal derivatives ............................... 98
5.3.3 Shifted delta coefficients ......................... 99
Concluding remarks ....................................... 100
6 Two-dimensional audio parameterization ................... 101
Introduction ............................................. 101
6.1 The audio spectrogram .................................... 102
6.2 The thresholded spectrogram .............................. 105
6.3 Morphological filtering of the thresholded spectrogram ... 107
6.4 MFCC'estimation with robust frames selection ............. 114
6.5 Points of interest-based features ........................ 121
6.6 Bag-of-instances audio descriptors ....................... 126
Concluding remarks ............................................ 135
7 Audio recognizers ........................................ 137
Introduction ............................................. 137
7.1 Overview of classification methods ....................... 137
7.2 One-species detectors .................................... 143
7.3 Single-label classification with a single model per
category ................................................. 148
7.4 Single-label classification with multi-classifier
schemes .................................................. 151
7.5 MIML classification schemes .............................. 154
Concluding remarks ....................................... 159
8 Application examples ..................................... 161
Introduction ............................................. 161
8.1 The ARBIMON project ...................................... 164
8.1.1 The ARBIMON Acoustics project ..................... 164
8.1.2 The ARBIMON II project ............................ 167
8.2 The AmiBio project ....................................... 170
8.3 The Pantanal Biomonitoring Project ....................... 178
8.4 The SABIOD project ....................................... 181
8.5 The ENTOMATIC project .................................... 182
8.6 The REMOSIS project ...................................... 184
Concluding remarks ....................................... 185
9 Useful resources ......................................... 189
Introduction ............................................. 189
9.1 Online audio data sets and resources ..................... 190
9.1.1 The xeno-canto repository ......................... 190
9.1.2 The Borror Laboratory of Bioacoustics Sound
Archive ........................................... 191
9.1.3 The Macaulay Library .............................. 191
9.1.4 The British Library Sound Archive ................. 191
9.1.5 Sound Library of the Wildlife Sound Recording
Society ........................................... 192
9.1.6 The BioAcoustica repository ....................... 192
9.1.7 The Animal Sound Archive at the Museum für
Naturkunde in Berlin .............................. 193
9.1.8 Australian National Wildlife Collection Sound
Archive ........................................... 193
9.1.9 The Western SoundScape Archive .................... 193
9.1.10 The DORSA Archive ................................. 194
9.1.11 The Nature Sounds Society ......................... 194
9.1.12 The Acoustics Ecology Institute ................... 194
9.1.13 Sound resources not listed here ................... 194
9.2 Software tools and services .............................. 195
9.3 Online information ....................................... 198
9.4 Scientific forums ........................................ 199
9.5 Technology evaluation campaigns .......................... 200
Concluding remarks ............................................ 203
Epilogue ...................................................... 205
References .................................................... 209
Index ......................................................... 221
| 
