Bird Species Image Identification using Transfer Learning

Bird Species Image Identification using Transfer Learning

Prof. N.V. Gawali1,  Tanmay V. Agavile2,  Aditya G.Dalvi3,  Shreyas S.Joshi4,

Computer Engineering Department,

PDEA’S College Of Engineering Manjari(BK), Pune-412307, India

Abstract- Nowadays, many new fellow Bird Watchers are having a hard time remembering all the bird species and identifying them. Also, common people and newly joined rescue team members are not able to identify bird species to be able to rescue them and treat them. They have to go through a hard way of identifying thick books like “Birds of Indian Subcontinent”. In this project paper, we evaluate and show the result of Deep Learning based AI Model which is useful for identifying birds using their images. The paper shows a Simple Web App which uses Transfer Learning, which is one of the best Deep Learning techniques, to identify images. We are going to use the InceptionV3 model made by google to get trained on the dataset of 325 different bird species having 1000 images per species with labels.

Keywords- Deep Learning, InceptionV3, Bird Identification, Computer Vision, ImageNet