Birds are present in various scenarios that appear in different shapes, sizes, and colors. It is estimated that there exist around 10,000 different species all over the world. Watching birds is a common practice but identifying their species requires bird knowledge. Human ability to recognize the birds through images is much easier as compared to audio classification, so bird species classification through images is in more trend and accurate too. Traditional machine learning and audio-based approaches are not suitable for bird classification. In this paper, we proposed a transfer learning-based approach for the classification of the bird dataset of 200 bird species. We performed numerous experiments with different deep learning models and the experimental results suggest that the proposed transfer learning-based approach significantly performs better than other state-of-the-art methods.
Publication: Advance computing
Publisher: Springer Link
Authors: Tejalal Choudhary, Shubham Gujar, Kruti Panchal, Sarvjeet, Vipul Mishra , Anurag Goswami
Keywords: Deep learning, Convolutional Neural Network, Transfer learning, Image classification