Bird Species Classification from Images Using Deep Learning

Learning about the birds improve the understanding of the world, and provides valuable information about the natural world. To assess the quality of the living environment, accurate data on the species of birds is important. Birds species classification and identification is a difficult task, even for expert biologists and ornithologists. The unavailability of experts, along with human limitations, further pose an upper limit on manual identification of birds and their species. Using an automated approach to identify birds and their species could be a significantly important idea in this scenario. In this paper, we evaluate several deep learning based models including SSD, YOLOv4 and YOLOv5 for birds species classification and identification. All the models are evaluated on publicly available CUB-200-2011 dataset. The YOLOv4 model outperforms the recent state-of-art methods with 95.43% accuracy, 93.94% precision, 94.34% recall and 94.27% F-1 score for 20 classes, along with 96.99% mAP score.