Deakin University
Browse

File(s) not publicly available

Detection of ground parrot vocalisation: A multiple instance learning approach

Version 2 2024-06-06, 02:45
Version 1 2018-06-05, 10:15
journal contribution
posted on 2024-06-06, 02:45 authored by Duc Thanh NguyenDuc Thanh Nguyen, PO Ogunbona, W Li, E Tasker, John YearwoodJohn Yearwood
Ground parrot vocalisation can be considered as an audio event. Test-based diverse density multiple instance learning (TB-DD-MIL) is proposed for detecting this event in audio files recorded in the field. The proposed method is motivated by the advantages of multiple instance learning from incomplete training data. Spectral features suitable for encoding the vocal source information of the ground parrot vocalization are also investigated. The proposed method was benchmarked against a dataset collected in various environmental conditions and an audio detection evaluation scheme is proposed. The evaluation includes a study on performance of the various vocal source features and comparison with other classification techniques. Experimental results indicated that the most appropriate feature to encode ground parrot calls is the spectral bandwidth and the proposed TB-DD-MIL method outperformed other existing classification methods.

History

Journal

Journal of the Acoustical Society of America

Volume

142

Article number

1281

Pagination

1281-1290

Location

United States

ISSN

0001-4966

eISSN

1520-8524

Language

English

Publication classification

E Conference publication, C1 Refereed article in a scholarly journal

Copyright notice

2017, Acoustical Society of America

Issue

3

Publisher

ACOUSTICAL SOC AMER AMER INST PHYSICS