Based on the work of Michael T. Johnson,Narjes Bozorg, Sidrah Liaqat, Neenu Jose, Patrick Conrey, Anthony Tamasi Original work here: https://github.com/UKYSpeechLab/ukybirddet
Required Software:
- Python version 3.6
- Tensorflow-gpu (tested on v1.8)
- Keras
- other packages and dependencies listed in requirements.txt
Project directory structure
For running this project on the DCASE challenge 2018 data, follow this directory structure: (unzip the provided compressed files in the main project directory)
- < project directory >/adaptation_files
- < project directory >/labels
- < project directory >/prediction
- < project directory >/trained_model
- < project directory >/workingfiles
The directory 'workingfiles' has four subdirectories that contain feature files.
- < project directory >/workingfiles/features_baseline (download link)
In order to reproduce the results of this submission, place the python files in the main project directory and run in the following order:
- birddet_baseline.py