Identifying bird species by their calls in Soundscapes
Metadatos
Mostrar el registro completo del ítemEditorial
Springer Nature
Materia
Multi-label classification Signal processing Domain mismatch Convolutional neural networks
Fecha
2023-03-20Referencia bibliográfica
K. Maclean and I. Triguero. Identifying bird species by their calls in Soundscapes. Applied Intelligence. [https://doi.org/10.1007/s10489-023-04486-8]
Patrocinador
Projects A-TIC-434-UGR20; PID2020-119478GB-I00; Maria Zambrano fellowship at the University of Granada; Universidad de Granada / CBUAResumen
In many real data science problems, it is common to encounter a domain mismatch between the training and testing datasets,
which means that solutions designed for one may not transfer well to the other due to their differences. An example of such
was in the BirdCLEF2021 Kaggle competition, where participants had to identify all bird species that could be heard in audio
recordings. Thus, multi-label classifiers, capable of coping with domain mismatch, were required. In addition, classifiers
needed to be resilient to a long-tailed (imbalanced) class distribution and weak labels. Throughout the competition, a diverse
range of solutions based on convolutional neural networks were proposed. However, it is unclear how different solution
components contribute to overall performance. In this work, we contextualise the problem with respect to the previously
existing literature, analysing and discussing the choices made by the different participants. We also propose a modular
solution architecture to empirically quantify the effects of different architectures. The results of this study provide insights
into which components worked well for this challenge.