I have my very first project starting on the 7th of January. This project involves using advanced sound software to develop auto species recognizers. I am aiming to develop a species recognition template for Barred Owls (Strix varia), with a minimal recognition accuracy of 80%.
This project is highly valuable, because it could allow scientists to digest sound data files much faster. When scientists record the calls of birds, frogs, and other animals, they have to go through the sound file(s) numerous times - and ideally, recordings are extensive. This can be A LOT of work! So, by developing auto-recognition templates, we can very quickly perform multiple runs through sound files and ID and quantify a target species, all while ignoring background audio.
Barred owls have a modest diversity of vocalizations, a collection of which can be heard at the following link:

No comments:
Post a Comment