Starting out a particle physics nerd back in school, I found my way to bioacoustics at the Centre for Biodiversity and Environment Research (CBER) whilst pursuing a BSc in Physics at University College London (UCL). I later earnt my MSc in Wildlife Conservation at Liverpool John Moores University (LJMU), before undertaking my PhD at St Andrews two years after. I am interested in all things sound! You can follow my non-academic explorations into acoustics on Instagram at @thenoisyplanet.
Towards an acoustic species classifier for North Atlantic odontocetes: A quantitative dive into the communication systems of Scotland’s toothed whales
- To investigate species-specific signaling in North Atlantic odontocete (toothed whale) species, en route to developing a model to classify species based on their vocalizations. Odontocetes produce a wide variety of sounds, ranging from echolocation ‘clicks’ to tonal ‘whistles’. Through vocal repertoire analysis and species classification on an extensive acoustic dataset, the study will help to answer the question of whether or not the properties of these sounds are distinguishable between the North Atlantic species, which has important implications for monitoring
- To quantitatively describe and compare the vocal repertoires (i.e., the range of call types) of North Atlantic odontocete species, filling gaps in our knowledge of the some of the lesser-studied species like the white-beaked dolphin (Lagenorhynchus albirostris) and long-finned pilot whale (Globicephala melas)
- To use deep learning to classify species by their vocalizations, comparing the performance of various models and data preparation techniques to provide a useful framework for future studies to build off
- To compare the performance of the deep learning classifiers with traditional machine learning classifiers, implementing the most accurate classifiers into open-source software for use by the scientific and wider community
- Email: firstname.lastname@example.org
- Twitter: @KleynTristan