Methods for classification of nocturnal migratory bird vocalizations using Pseudo Wigner-Ville Transform

Abstract

Many species of birds in Americas vocalize during nocturnal migration flights. Acoustic detection and classification of the calls show potential for study of the natural history of these migrant birds. In particular, information about the species' composition and number of birds involved in migration movements may be obtainable through acoustic techniques. Other methods such as radar monitoring may have capability only to assess the number, but not the composition. Mel Frequency Cepstral Coefficients-Gaussian Mixture Model-based methods (MFCC-GMM), Mel Frequency Cepstral Coefficients-Hidden Markov Model-based methods (MFCC-HMM) and spectrogram correlation-based methods have been proposed to automate the recognition/classification of the nocturnal flight calls. Here we investigate the choice of Pseudo Wigner-Ville Transform (PWVT) on MFCC-HMM-based classifier and correlation-based classifier performance. We use a collection of recordings of nocturnal flight calls of several species of thrushes and other bird species with similar calls to evaluate and compare classifiers.