Our team is dedicated to advancing machine learning techniques for the study of Himalayan bird songs, with a special focus on Hume’s Leaf Warbler. This research combines cutting-edge deep learning methods and acoustic analysis to answer key questions about species identification, distribution, and behavior. By leveraging and extending tools like the BirdNET model, we aim to refine bird call detection and investigate the unique vocalizations of Hume’s Leaf Warbler, particularly the enigmatic “buzz” vocalization.
Our work seeks to uncover:
- Behavioral hypotheses about the buzz vocalization, such as its role in mate attraction early in the breeding season.
- How well BirdNET performs in identifying birds in our experimental site.
- How to develop custom models to detect the warbler’s calls and specific vocalizations.
- Insights into species distribution patterns across elevational gradients and seasonal shifts.
With a rich dataset of acoustic recordings from the Himalayan breeding season and advanced techniques like spectrogram analysis, temporal modeling, and clustering algorithms, we aim to contribute to ecological understanding and machine learning innovation.
Join us as we explore the intersection of technology, biology, and conservation.