Feature Extraction For Western Ghats Birds

Abstract

The research paper focuses on developing an automated model for extracting different kinds of features from bird’s images and prepares a bench marking dataset which will help to evaluate the performance of the classification algorithms in an effective manner. This proposed work has been implemented using Visual Studio .net framework in which C#.Net is the front end tool and SQL Server is the back end tool. The user has to upload image in the format of jpeg or png. This tool automatically analyzes and extracts prominent features of such as wing type, wing pattern, shape, edge tail, legs from the Western Ghats bird image using Convolution Neural Network (CNN). The extracted features of an image will be stored in the database and to retrieve those features of an image, CNN uses small square bounding box of image pixels. This tool helps the user to extract the entire prominent feature which plays a vital role in classification process.