Detection and Segmentation of Bird Song in Noisy Environments

Abstract

Recent work in machine learning concerns the detection and identification of bird species from audio recordings of their vocalizations. Such analysis can yield valuable ecological information concerning the activity and distribution of species in the wild. Current species-identification methods require individual syllables of bird audio as input, but field-collected audio contains noise and simultaneous vocalizations. This thesis presents two supervised learning methods for identifying and separating individual syllables of bird vocalizations from field-recorded audio. The segments output by this process can be input into species classification algorithms, to yield useful ecological data.