AEIN-An Intelligent Computational Technique for Biometric Based Individual Yorkshire Pig Identification Using Auricular Vein

Abstract

This article introduces a novel technology called “AEIN” that has been suggested for the unique identification of individual pigs based on their auricular vein pattern. The aim is to address the limitations of conventional identification methods, which are known for their unreliability, inaccuracy, and susceptibility to manipulation, while also promoting the concept of intelligent livestock management. The ear images of pigs were obtained from a certified farm, processed according to established protocols, and subjected to feature extraction to create templates for subsequent matching within the same class and across different classes using various distance metrics such as Euclidean, Manhattan, Minkowski, and Hamming distances. Specifically, the points of branching in the vein pattern were utilized as features for template creation. By carefully analyzing each distance metric, a threshold level was established, with the average distance set at 20 for Manhattan distance and 40 for Minkowski, Euclidean, and Hamming distances, respectively. If the calculated matching distance falls below the threshold, the pig is successfully identified; otherwise, it is considered a different individual. The Euclidean distance metric demonstrated the highest accuracy in identification among all four metrics in the conducted experiments. A total of 54 pigs were included in the study, revealing that the “AEIN” technology achieved a remarkable accuracy rate of 98.18% when employing the Euclidean distance metric.