Stacked Convolutional and Recurrent Neural Networks for Bird Audio Detection
This work addresses bird audio detection for ecological monitoring, with incremental improvements in robustness to unseen data.
The paper tackles bird call detection in audio segments by proposing stacked convolutional and recurrent neural networks with data augmentation and domain adaptation methods, achieving 95.5% AUC on development data and 88.1% on unseen evaluation data.
This paper studies the detection of bird calls in audio segments using stacked convolutional and recurrent neural networks. Data augmentation by blocks mixing and domain adaptation using a novel method of test mixing are proposed and evaluated in regard to making the method robust to unseen data. The contributions of two kinds of acoustic features (dominant frequency and log mel-band energy) and their combinations are studied in the context of bird audio detection. Our best achieved AUC measure on five cross-validations of the development data is 95.5% and 88.1% on the unseen evaluation data.