A reliable recognition System for Ear recognition

Abstract

In the last decade, ear recognition has been the subject matter of several researches in the field of computer vision due to its interesting applications, especially the forensic applications. In this work, we develop a new system for human ear print recognition based on local texture descriptors. In particular, we employ three well-known descriptors namely Local Binary Patterns, Grey-Level Co-Occurrence matrix (GLCM) and BinariazeStatistical Image Features (BSIF), where K-Nearest Neighbour classifier is used to perform classification. We carried out experiments on the IIT-Delhi-I dataset which is made up of 493 images from 125 subjects, where an overall accuracy of % is reached.