Paper Key : IRJ************883
Author: Vidyashree K P
Date Published: 05 Jul 2024
Abstract
This research presents a new approach to identifying bird species using audio data processing and convolutional neural networks (CNN). Following their collection, recordings of birdsong were processed to extract crucial elements known as mel-frequency ceptral coefficient(MFCC). These characteristics were then applieds to train a CNN model designed to accurately classify different species of birds. Across a variety of datasets, our methodology greatly increased classification accuracy when compared to traditional methods. The results illustrate the high degree of coherence between processing audio signal and deep learning for bioacoustic applications, providing new opportunities for automated wildlife monitoring and conservation efforts.