CityNet - Deep Learning Tools for Urban Ecoacoustic Assessment
This collection is shared privately
Two expert-annotated urban audio datasets presented in Fairbrass, A.J., Firman, M. et al. (2018) 'CityNet ‐ Deep Learning Tools for Urban Ecoacoustic Assessment'. MEE, https://doi.org/10.1111/2041-210X.13114 are
made freely available here.
CityNet is a machine learning system for estimating the level of biotic and anthropogenic sound at each moment in time in an audio file.
The system has been trained and validated on human-labelled audio files captured from green spaces around London.
CityNet comprises a neural network classifier, which operates on audio spectrograms to produce a measure of biotic or anthropogenic activity level.
AudioTagger software (https://github.com/groakat/AudioTagger) is required to use the CitySounds2017 annotations.
REFERENCES
COLLECTION SIZE
3.08 GB