Use this URL to cite or link to this record in EThOS:
Title: Automated classification of humpback whale (Megaptera novaeangliae) songs using hidden Markov models
Author: Pace, Federica
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Humpback whales songs have been widely investigated in the past few decades. This study proposes a new approach for the classification of the calls detected in the songs with the use of Hidden Markov Models (HMMs). HMMs have been used once before for such task but in an unsupervised algorithm with promising results. Here HMMs were trained and two models were employed to classify the calls into their component units and subunits. The results show that classification of humpback whale songs from one year to another is possible even with limited training. The classification is fully automated apart from the labelling of the training set and the input of the initial HMM prototype models. Two different models for the song structure are considered: one based on song units and one based on subunits. The latter model is shown to achieve better recognition results with a reduced need for updating when applied to a variety of recordings from different years and different geographic locations.
Supervisor: White, Paul Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: QC Physics ; QL Zoology