Paper Title
ENDANGERED IMAGE BASED BIRD SPECIES PREDICTION USING CNN AND TRANSFER LEARNING

Abstract
This Birds are a widely diverse class of warm-blooded creatures, encompassing approximately 10,000 living species with a myriad of characteristics and appearances. While birdwatching is a popular pastime, accurate species identification often necessitates specialized knowledge in the field of ornithology. To address this challenge, we present a convolutional neural network (CNN)-based automated model capable of distinguishing between various bird species using image data. Our model was trained on a combination of the Caltech-UCSD Birds 200 [CUB-200-2011] and BIRDS 525 Species datasets available on Kaggle. The architecture of our deep neural network was meticulously designed to extract meaningful features from bird images for accurate categorization. We systematically explored different hyperparameters and augmentation techniques to enhance model performance. Our experimental results demonstrate that the proposed model achieves a commendable accuracy of 98.76% on the test dataset. Moreover, we underscore the significance of leveraging technology for the conservation and management of endangered bird populations, highlighting the potential of convolutional neural networks in bird species identification. In summary, our methodology offers a valuable tool for population monitoring and conservation efforts, with the potential for further improvements in accuracy and extension to encompass a broader range of bird species. Keywords - Bird species; Machine Learning; Convolutional Neural Networks; Ornithology.