iBon: A Web Application for Aerial Fauna Identification and Counting Using Machine Learning

With the alarming decline in aerial fauna populations worldwide, the need for timely and accurate tools to monitor species trends and support conservation strategies has become critical. This paper aims to develop and evaluate iBon, a web-based application that provides automated bird identification and counting using advanced machine learning models. Traditional methods like manual observation are time-consuming, prone to observer bias, and inconsistent across datasets. iBon addresses these challenges by employing a Convolutional Neural Network (CNN) for bird identification, achieving 94% accuracy across 17 datasets, with performance boosted through a pre-trained MobileNet feature extractor. The system integrates YOLOv8, a fast and accurate object detection model for bird counting. Both models are assessed using accuracy, F1-score, and robustness to dataset variations. iBon delivers a reliable and user-friendly platform that empowers researchers, conservationists, and citizen scientists with efficient tools for biodiversity monitoring and data-driven decision-making.