Fusion of shape of the ear and tragus – A unique feature extraction method for ear authentication system

Abstract

In the current scenario to access the data securely from wireless network, e-commerce and internet have become a great concern. As the technologies become commercialized in high markets growth, there is a need for effective and secure implementation to authenticate the identity of the people. Several biometric techniques that exploit physiological and behavioral traits of people have been developed for verification and identification of the individuals to make the system more reliable and secure. In this paper a unique method is used to extract the tragus feature of the ear. The proposed method used is an enhanced edge detection method. A line is drawn connecting the maximum coordinate, midpoint and the minimum coordinate. The mid region of this line gives the tragus of the ear. The ear feature tragus is taken because it is almost free from occlusions. In order to develop an efficient ear authentication system the feature level fusion is applied. The shape of the ear is also extracted. Then the tragus and the shape of the ear are combined together to form the fused template. The authentication rate of the system with the features of tragus, shape of the ear and the fused template is calculated using the Hamming distance technique and Euclidean distance method. The experimental result is carried out on two databases of IIT Delhi ear database and Mathematical Analysis of Images (AMI) ear database including occlusion. The proposed system is to prove that the accuracy of the ear biometric authentication system using Hamming distance technique in fused feature is better than the individual feature of tragus and the ear shape. When a single feature tragus is considered tragus is better than the ear shape as it is a unique feature of the ear.