Abstract
One significant challenge to biodiversity assessment and conservation is persistent gaps in species diversity knowledge in Earth’s most biodiverse areas. Monitoring devices that utilize species-specific advertisement calls show promise in overcoming challenges associated with lagging frog species discovery rates. However, these devices generate data at paces faster than it can be analyzed. As such, automated platforms capable of efficient data processing and accurate species-level identification are at a premium. In addressing this gap, we used TensorFlow Inception v3 to design a robust, automated species identification system for 41 Philippine frog species (genus Platymantis), utilizing single-note audio spectrograms. With this model, we explored two concepts: (1) performance of our deep-learning model in discriminating closely-related frog species based on images representing advertisement call notes, and (2) the potential of this platform to accelerate new species discovery. TensorFlow identified species with a ~ 94% overall correct identification rate. Incorporating distributional data increased the overall identification rate to ~ 99%. In applying TensorFlow to a dataset that included undescribed species in addition to known species, our model was able to differentiate undescribed species through variation in “certainty” rate; the overall certainty rate for undescribed species was 65.5% versus 83.6% for described species. This indicates that, in addition to discriminating recognized frog species, our model has the potential to flag possible new species. As such, this work represents a proof-of-concept for automated, accelerated detection of novel species using acoustic mate-recognition signals, that can be applied to other groups characterized by vibrational cues, seismic signals, and vibrational mate-recognition.
Similar content being viewed by others
Data availability
Data are publicly available via Cornell University’s Laboratory of Ornithology and Macaulay Library of Natural Sounds (https://www.macaulaylibrary.org).
References
Abadi M et al. (2016) Tensorflow: a system for large-scale machine learning. Paper presented at the 12th USENIX Symposium on Operating Systems Design and Implementation, Savannah, USA,
Aide TM, Corrada-Bravo C, Campos-Cerqueira M, Milan C, Vega G, Alvarez R (2013) Real-time bioacoustics monitoring and automated species identification. PeerJ 1:e103
Araya-Salas M, Smith-Vidaurre G (2017) warbleR: an R package to streamline analysis of animal acoustic signals. Methods Ecol Evol 8:184–191
Brabant R, Laurent Y, Dolap U, Degraer S, Poerink BJ (2018) Comparing the results of four widely used automated bat identification software programs to identify nine bat species in coastal Western Europe Belgian. J Zool 148:119–128
Brown RM, Gonzalez JC (2007) A new forest frog of the genus Platymantis (Amphibia: Anura: Ranidae) from the Bicol Peninsula of Luzon Island, Philippines. Copeia 2007:251–266
Brown R, Stuart B (2012) Patterns of biodiversity discovery through time: an historical analysis of amphibian species discoveries in the Southeast Asian mainland and island archipelagos. In: Gower D et al (eds) Biotic evolution and environmental change in Southeast Asia. Cambridge University Press, Cambridge, pp 348–389
Brown WC, Alcala A, Diesmos AC (1997a) A new species of the genus Platymantis (Amphibia: Ranidae) from Luzon Island, Philippines. Proc Biol Soc Wash 110:18–23
Brown WC, Alcala A, Diesmos AC, Alcala E (1997b) Species of the guentheri group of Platymantis (Amphibia: Ranidae) from the Philippines, with descriptions of four new species. Proc Calif Acad Sci 50:1–20
Brown W, Alcala A, Ong P, Diesmos A (1999a) A new species of Platymantis (Amphibia: Ranidae) from the Sierra Madre Mountains, Luzon Island, Philippines. Proc Biol Soc Wash 112:510–514
Brown WC, Alcala AC, Diesmos AC (1999b) Four new species of the genus Platymantis (Amphibia: Ranidae) from Luzon Island, Philippines. Proc Calif Acad Sci 51:449–460
Brown RM et al (2013) Evolutionary processes of diversification in a model island archipelago. Annu Rev Ecol Evol Syst 44:411–435
Brown RM, De Layola LA, Lorenzo A, Diesmos MLL, Diesmos AC (2015a) A new species of limestone karst inhabiting forest frog, genus Platymantis (Amphibia: Anura: Ceratobatrachidae: subgenus Lupacolus) from southern Luzon Island, Philippines. Zootaxa 4048:191–210
Brown RM, Siler CD, Richards SJ, Diesmos AC, Cannatella DC (2015b) Multilocus phylogeny and a new classification for Southeast Asian and Melanesian forest frogs (family Ceratobatrachidae). Zool J Linnean Soc 174:130–168
Bryan JE, Shearman PL, Asner GP, Knapp DE, Aoro G, Lokes B (2013) Extreme differences in forest degradation in Borneo: coamparing practices in Sarawak, Sabah, and Brunei. PLoS ONE 8:e69679
Chan KO, Brown RM (2017) Did true frogs ‘dispersify’? Biol Lett 13:20170299
Chen Z, Wiens JJ (2020) The origins of acoustic communication in vertebrates. Nat Commun 11:1–8
Diesmos AC et al (2015) Amphibians of the Philippines, part I: checklist of the species. Proc Calif Acad Sci 62:457–539
Feinberg JA et al (2014) Cryptic diversity in metropolis: confirmation of a new leopard frog species (Anura: Ranidae) from New York City and surrounding Atlantic coast regions. PLoS ONE 9:e108213
Gaston KJ, O’Neill MA (2004) Automated species identification: why not? Philos Trans R Soc Lond B 359:655–667
Gerhardt HC (1978) Mating call recognition in the green treefrog (Hyla cinerea): the significance of some fine-temporal properties. J Exp Biol 74:59–73
Gerhardt HC (1994) The evolution of vocalization in frogs and toads. Annu Rev Ecol Syst 25:293–324
González-del-Pliego P, Freckleton RP, Edwards DP, Koo MS, Scheffers BR, Pyron RA, Jetz W (2019) Phylogenetic and trait-based prediction of extinction risk for data-deficient amphibians. Curr Biol 29(1557–1563):e1553
Guirado E, Tabik S, Rivas ML, Alcaraz-Segura D, Herrera F (2019) Whale counting in satellite and aerial images with deep learning. Sci Rep 9:14259
Gurgel-Gonçalves R et al (2017) Automated identification of insect vectors of chagas disease in Brazil and Mexico: the virtual vector. Lab PeerJ 5:e3040
Holmgren J, Persson Å, Söderman U (2008) Species identification of individual trees by combining high resolution LiDAR data with multi-spectral images. Int J Remote Sens 29:1537–1552
Inger RF (1954) Systematics and zoogeography of Philippine Amphibia. Fieldiana Zool 33:181–531
Khalighifar A, Komp E, Ramsey JM, Gurgel-Gonçalves R, Peterson AT (2019) Deep learning algorithms improve automated identification of chagas disease vectors. J Med Entomol 56:1404–1410
Klein DJ, McKown MW, Tershy BR (2015) Deep learning for large scale biodiversity monitoring. Paper presented at the Bloomberg Data for Good Exchange Conference, New York City, USA
Kohavi R (1995) A study of cross-validation and bootstrap for accuracy estimation and model selection. In: International Joint Conferences on Artificial Intelligence, Montreal, Canada, pp. 1137–1145
Kumar N, Belhumeur PN, Biswas A, Jacobs DW, Kress WJ, Lopez IC, Soares JV (2012) Leafsnap: A computer vision system for automatic plant species identification. In: European Conference on Computer Vision, Florence, Italy, pp. 502–516
MacLeod N, Benfield M, Culverhouse P (2010) Time to automate identification. Nature 467:154–155
Molinaro AM, Simon R, Pfeiffer RM (2005) Prediction error estimation: a comparison of resampling methods. Bioinformatics 21:3301–3307
Narins PM, Capranica RR (1977) An automated technique for analysis of temporal features in animal vocalizations. Anim Behav 25:615–621
Philippe J, Felipe L, Celio F (2017) The use of bioacoustics in anuran taxonomy: theory, terminology, methods and recommendations for best practice. Zootaxa 4251:1–124
Ramcharan A, Baranowski K, McCloskey P, Ahmed B, Legg J, Hughes DP (2017) Deep learning for image-based cassava disease detection. Front Plant Sci 8:1852–1859
Russakovsky O et al (2015) ImageNet large scale visual recognition challenge. Int J Comput Vision 115:211–252
Rzanny M, Seeland M, Wäldchen J, Mäder P (2017) Acquiring and preprocessing leaf images for automated plant identification: understanding the tradeoff between effort and information gain. Plant Methods 13:1–11
Scheffers BR, Brunner RM, Ramirez SD, Shoo LP, Diesmos A, Williams SE (2013) Thermal buffering of microhabitats is a critical factor mediating warming vulnerability of frogs in the Philippine biodiversity hotspot. Biotropica 45:628–635
Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Netw 61:85–117
Siler CD, Linkem CW, Diesmos AC, Alcala AC (2007) A new species of Platymantis (Amphibia: Anura: Ranidae) from Panay Island, Philippines. Herpetologica 63:351–364
Siler CD, Diesmos AC, Linkem CW, Diesmos ML, Brown RM (2010) A new species of limestone-forest frog, genus Platymantis (Amphibia: Anura: Ceratobatrachidae) from central Luzon Island, Philippines. Zootaxa 2482:49–63
Smith SW et al (2019) A deep neural network learning algorithm outperforms a conventional algorithm for emergency department electrocardiogram interpretation. J Electrocardiol 52:88–95
Sueur J, Aubin T, Simonis C (2008) Seewave, a free modular tool for sound analysis and synthesis. Bioacoustics 18:213–226
Sugai LSM, Silva TSF, Ribeiro JW Jr, Llusia D (2019) Terrestrial passive acoustic monitoring: review and perspectives. Bioscience 69:15–25
Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z (2016) Rethinking the inception architecture for computer vision. Paper presented at the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA,
Tapley B, Michaels CJ, Gumbs R, Böhm M, Luedtke J, Pearce-Kelly P, Rowley JJ (2018) The disparity between species description and conservation assessment: a case study in taxa with high rates of species discovery. Biol Conserv 220:209–214
Taylor EH (1923) Addition to the herpetological fauna of the Philippine Islands III Philippine. J Sci 22:515–557
Tonini JFR, Provete DB, Maciel NM, Morais AR, Goutte S, Toledo LF, Pyron RA (2020) Allometric escape from acoustic constraints is rare for frog calls. Ecol Evol 10:1–10
Vieites DR, Wollenberg KC, Andreone F, Köhler J, Glaw F, Vences M (2009) Vast underestimation of Madagascar’s biodiversity evidenced by an integrative amphibian inventory. Proc Natl Acad Sci USA 106:8267–8272
Vignal C, Kelley D (2007) Significance of temporal and spectral acoustic cues for sexual recognition in Xenopus laevis. Proc R Soc B 274:479–488
Villa AG, Salazar A, Vargas F (2017) Towards automatic wild animal monitoring: identification of animal species in camera-trap images using very deep convolutional neural networks. Ecol Inform 41:24–32
Wells KD (2010) The ecology and behavior of amphibians. University of Chicago Press, Chicago
Wells KD, Schwartz JJ (2007) The behavioral ecology of anuran communication. Hearing and sound communication in amphibians, vol 28. Springer, New York, pp 44–86
Wimmer J, Towsey M, Roe P, Williamson I (2013) Sampling environmental acoustic recordings to determine bird species richness. Ecol Appl 23:1419–1428
Zhao Z, Zhang S, Xu Z, Bellisario K, Dai N, Omrani H, Pijanowski BC (2017) Automated bird acoustic event detection and robust species classification. Ecol Inform 39:99–108
Acknowledgements
Philippine Platymantis frog calls were collected with support from the U. S. National Science Foundation’s former Doctoral Dissertation Improvement Grant (DEB 0073199; 2001–2003) and a Biotic Surveys and Inventories Grant (DEB 0743491; 2008–2012). Collection and co-curated voucher specimens and their associated digital media specimens, archival digitization, data verification, and online serving of digital media specimens was made possible by a NSF Thematic Collections Network (TCN) program Grant (DEB 1304585; 2013–2018). Recent extended specimen collection and curation has been supported by NSF DEB 1654388 and 1557053, with further support from the KU Biodiversity Institute’s Rudkin Research Exploration (REX) Fund and KU College of Liberal Arts and Sciences Docking Scholar Fund. We thank A. Diesmos, J. Fernandez, C. Siler, and C. Meneses for help recording frogs.
Funding
No funding was used for this research.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
All authors declare that they have no conflicts of interest.
Additional information
Communicated by Dirk Sven Schmeller.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Below is the link to the Supplementary information.
Rights and permissions
About this article
Cite this article
Khalighifar, A., Brown, R.M., Goyes Vallejos, J. et al. Deep learning improves acoustic biodiversity monitoring and new candidate forest frog species identification (genus Platymantis) in the Philippines. Biodivers Conserv 30, 643–657 (2021). https://doi.org/10.1007/s10531-020-02107-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10531-020-02107-1