Content-based audio classification system for bird sounds
Birds are very important to the ecosystem and an agent in promoting biodiversity. Their vocalizations consist of songs and calls, and are used to communicate, i.e., mating calls, warning calls, etc. This paper aims to automatically classify bird sounds from five native Malaysian birds – the Rhinocer...
Saved in:
Main Authors: | , , , , |
---|---|
Format: | Article |
Published: |
Semarak Ilmu Publishing
2024
|
Online Access: | http://psasir.upm.edu.my/id/eprint/105846/ https://semarakilmu.com.my/journals/index.php/applied_sciences_eng_tech/article/view/3786 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my.upm.eprints.105846 |
---|---|
record_format |
eprints |
spelling |
my.upm.eprints.1058462024-05-08T23:39:24Z http://psasir.upm.edu.my/id/eprint/105846/ Content-based audio classification system for bird sounds Norowi, Noris Mohd Anuar, Nur Asilah Mustaffa, Mas Rina Hussin, Masnida Nasharuddin, Nurul Amelina Birds are very important to the ecosystem and an agent in promoting biodiversity. Their vocalizations consist of songs and calls, and are used to communicate, i.e., mating calls, warning calls, etc. This paper aims to automatically classify bird sounds from five native Malaysian birds – the Rhinoceros Hornbill, the Black and Yellow Broadbill, the Common Myna, the Malayan Banded Pitta, and the Crested Serpent Eagle. In the initial experiment, the factors that affect the classification accuracy was studied. Results from the initial became the basis of the development of the MyBird5ounds system, a PC-based standalone system that was build using MATLAB. By applying different features combinations, the classification results differed, and the combination that resulted in the improvement of the classification results were identified. The contribution of this paper lies in the small-scale study that compares the performance of manual bird sounds classification by humans and the automatic classification from MyBird5ounds. 80 classification accuracy was achieved when the optimized parameters were applied – almost twice that achieved in manual classification by trained humans with no prior background in bird watching. This suggests that such a system is beneficial in aiding classification of birds using content-based audio classification methods. Semarak Ilmu Publishing 2024 Article PeerReviewed Norowi, Noris Mohd and Anuar, Nur Asilah and Mustaffa, Mas Rina and Hussin, Masnida and Nasharuddin, Nurul Amelina (2024) Content-based audio classification system for bird sounds. Journal of Advanced Research in Applied Sciences and Engineering Technology, 33 (3). pp. 307-318. ISSN 2462-1943 https://semarakilmu.com.my/journals/index.php/applied_sciences_eng_tech/article/view/3786 10.37934/araset.33.3.307318 |
institution |
Universiti Putra Malaysia |
building |
UPM Library |
collection |
Institutional Repository |
continent |
Asia |
country |
Malaysia |
content_provider |
Universiti Putra Malaysia |
content_source |
UPM Institutional Repository |
url_provider |
http://psasir.upm.edu.my/ |
description |
Birds are very important to the ecosystem and an agent in promoting biodiversity. Their vocalizations consist of songs and calls, and are used to communicate, i.e., mating calls, warning calls, etc. This paper aims to automatically classify bird sounds from five native Malaysian birds – the Rhinoceros Hornbill, the Black and Yellow Broadbill, the Common Myna, the Malayan Banded Pitta, and the Crested Serpent Eagle. In the initial experiment, the factors that affect the classification accuracy was studied. Results from the initial became the basis of the development of the MyBird5ounds system, a PC-based standalone system that was build using MATLAB. By applying different features combinations, the classification results differed, and the combination that resulted in the improvement of the classification results were identified. The contribution of this paper lies in the small-scale study that compares the performance of manual bird sounds classification by humans and the automatic classification from MyBird5ounds. 80 classification accuracy was achieved when the optimized parameters were applied – almost twice that achieved in manual classification by trained humans with no prior background in bird watching. This suggests that such a system is beneficial in aiding classification of birds using content-based audio classification methods. |
format |
Article |
author |
Norowi, Noris Mohd Anuar, Nur Asilah Mustaffa, Mas Rina Hussin, Masnida Nasharuddin, Nurul Amelina |
spellingShingle |
Norowi, Noris Mohd Anuar, Nur Asilah Mustaffa, Mas Rina Hussin, Masnida Nasharuddin, Nurul Amelina Content-based audio classification system for bird sounds |
author_facet |
Norowi, Noris Mohd Anuar, Nur Asilah Mustaffa, Mas Rina Hussin, Masnida Nasharuddin, Nurul Amelina |
author_sort |
Norowi, Noris Mohd |
title |
Content-based audio classification system for bird sounds |
title_short |
Content-based audio classification system for bird sounds |
title_full |
Content-based audio classification system for bird sounds |
title_fullStr |
Content-based audio classification system for bird sounds |
title_full_unstemmed |
Content-based audio classification system for bird sounds |
title_sort |
content-based audio classification system for bird sounds |
publisher |
Semarak Ilmu Publishing |
publishDate |
2024 |
url |
http://psasir.upm.edu.my/id/eprint/105846/ https://semarakilmu.com.my/journals/index.php/applied_sciences_eng_tech/article/view/3786 |
_version_ |
1800093788380594176 |
score |
13.188404 |