Data Science & Analytics

BIOACOUSTIC CLASSIFICATION OF AVIAN CALLS FROM RAW SOUND WAVEFORMS 

Francisco Bravo, PhD Student Central Queensland University

Wednesday November 10th 5-6 PM

Zoom

TALKING POINTS

• The use of autonomous sound recordings in fauna surveys and their challenges
• Traditional sound processing techniques (sound transformations), their use in deep learning and possible biases when applied to bioacoustic sounds  
• SincNet, a new approach applying machine learning directly to the sound 
• Testing SincNet in a public bioacoustic dataset and comparative results 

BIO

Born in Spain, I completed a Masters Degree in Biology at the university of Salamanca (1989-1994). I moved to Australia at the turn of the millennium, where I have worked primarily as an environmental consultant. I have been fortunate to learn about and experience firsthand, the fascinating Australian flora and fauna. My work has involved surveying animals and plants, mapping vegetation, assessing wetlands, trapping and handling marsupials and of course recording and analysing animal sounds, particularly bats. In 2020, I enrolled in a PhD part-time with the University of Central Queensland trying to advance the use of machine learning in bioacoustic sound processing. As part of my research, we are collaborating with Queensland Parks authority in the search for a subspecies of parrot (the Coxen’s fig-parrot) that has no recent confirmed sightings.

Download Flyer Here