RetIS: Unique Identification System of Goats through Retinal Analysis

Abstract

Developing an accurate, reliable and tamper-proof unique identification system for identifying individual animals is one of the most important but challenging tasks in intelligent livestock management. Currently, unique identification for animals is done using methods like stenciling, ear tagging, ear notching, branding, etc. which create opportunities for the poachers, middle men and animal thieves to manipulate the tags and transfer animals across international borders, claim false insurance, poach for their meat, bones, etc. illegally. This becomes a huge headache for the livestock owners. In this paper, a novel biometric unique identification system called “RetIS” has been developed based on goat retinal image which is unique, tamper-proof, reliable and accurate. The identification system has been developed based on the origination and distribution pattern of the blood vessels through a circular ring around the optic disc, which remains constant for the entire lifetime of the goat. Goat retinal images have been captured using a handheld fundus camera. The captured images are then segmented, normalized and encoded for the creation of templates, which are eventually stored in a database for matching with query templates based on hamming distance criteria. In matching, it has been observed that if the hamming distance between stored and query template is less than equal to 0.37 (matching >= 63%), it can be confirmed that the stored and query template belongs to the same individual goat providing its unique identification, otherwise the templates belong to two separate individual goats. This model has been tested with more than 200 retinal images obtained from twelve different goats producing 99% accuracy. The performance of this proposed model has been compared with other animal identification technologies and is found to be the most accurate and precise.