Multi-Label Classifier Chains for Bird Sound
This work addresses the challenge of accurately identifying multiple bird species in audio recordings, which is important for ecological monitoring and citizen science, but it is incremental as it builds on existing classifier chain techniques.
The authors tackled the problem of multi-label classification in birdsong data, where multiple bird species vocalize simultaneously, by proposing an ensemble of classifier chains with a histogram-of-segments representation. Their method generally outperformed binary relevance and showed mixed results compared to existing multi-instance multi-label learning algorithms on two real-world datasets.
Bird sound data collected with unattended microphones for automatic surveys, or mobile devices for citizen science, typically contain multiple simultaneously vocalizing birds of different species. However, few works have considered the multi-label structure in birdsong. We propose to use an ensemble of classifier chains combined with a histogram-of-segments representation for multi-label classification of birdsong. The proposed method is compared with binary relevance and three multi-instance multi-label learning (MIML) algorithms from prior work (which focus more on structure in the sound, and less on structure in the label sets). Experiments are conducted on two real-world birdsong datasets, and show that the proposed method usually outperforms binary relevance (using the same features and base-classifier), and is better in some cases and worse in others compared to the MIML algorithms.