Bird Call Recognition using Acoustic based Feature Selection approach in Machine Learning

The recognition of bird calls has been a challenging task in the field of bioacoustics. With the advancement of machine learning algorithms, automatic bird call recognition has become an active research area. In this paper, acoustic feature selection is used which involves the extraction of relevant features from audio recordings of bird calls. Then different classification algorithms are explored for the recognition of the bird calls. Evaluation of the approach is done on a dataset consisting of recordings from multiple bird species and compared it with other state-of-the-art machine learning algorithms. This research has the potential to contribute to the conservation of bird species and their habitats by enabling the efficient monitoring of bird populations. The data is classified into four different classes (namely Astfly, Bulori, Warvir and Woothr). All of which is found in the depths of Amazon Rainforest.