Make Your Own Recogniser

As part of the QUT Ecoacoustics Symposium 2022 we ran a one day workshop for building automated call recognisers. The goal of this workshop was for each participant to leave with a working call recogniser for a single species-call of their choice. It covered some basic theory and practice of building a call recogniser using a convolutional neural network.

Our facilitators were Dr Philip Eichinski, and Dr Lance De Vine.

Dr Eichinski is a postdoctoral researcher and research software engineer specialising in machine learning for species call recognition. See his most recent publication: “A Convolutional Neural Network Bird Species Recognizer Built From Little Data by Iteratively Training, Detecting, and Labeling”.

Dr De Vine is a researcher in data science and machine learning with applications to multi-disciplinary research. Lance has been working on building species recognisers for five threatened species in the Gondwana Rainforests of Australia WHA, as well as the Glossy Black Cockatoo.

Below are the requirements and practical steps for you to make your own automated call recogniser:

• Requirements
• Practical Steps