Bird chant’s recognition based on principal components analysis of the spectrogram
This paper presents a method for automatic bird song recognition. This method is based on the principal components analysis (PCA) of bird chant’s spectrograms, by using eigenfaces. For the database confection, we used a detection method for acoustic activity applying a moving average filter. We asse...
Autores Principales: | González, Jessica, Padrón, Jorge, Barbero, Iván, Custodio, Luis, Fernando Merchán, Fernando Merchán |
---|---|
Formato: | Artículo |
Idioma: | Español |
Publicado: |
Universidad Tecnológica de Panamá
11/1
|
Acceso en línea: |
https://revistas.utp.ac.pa/index.php/ric/article/view/2398 https://ridda2.utp.ac.pa/handle/123456789/10101 |
Sumario: |
This paper presents a method for automatic bird song recognition. This method is based on the principal components analysis (PCA) of bird chant’s spectrograms, by using eigenfaces. For the database confection, we used a detection method for acoustic activity applying a moving average filter. We assembled a database of three chants of different birds: Formicivora Grisea, Harpia Harpyja and Reinita Protonotaria. The obtained results show the effectivity of the proposed method, obtaining a recognition rate of 100%. |
---|