Detection and Identification of Animals in Wild Life Sanctuaries using Convolutional Neural Network

Abstract

Comprehensive, Precise and real time data regarding the position and characteristics of animals is necessary for safeguarding visitors inside a wildlife sanctuary. Investigations are made on the capability for automated, unambiguous and economical collection of data that are useful to perform rescue operations within the sanctuary because of the absence of other communicational sources. Web camera enables collection of photos relating to wildlife economically, conservatively as well as regularly. Extraction of information from such photos is costly, slow and requires human intervention. The proposed system demonstrates the automatic extraction of such data using Convolutional Neural Network (CNN). Deep CNN is trained for a set of images available in a wildlife dataset.