Security and Privacy in Communication Networks. SecureComm 2017 International Workshops, ATCS and SePrIoT, Niagara Falls, ON, Canada, October 22–25, 2017, Proceedings

Research Article

Fast and Robust Biometric Authentication Scheme Using Human Ear

Download
184 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-78816-6_11,
        author={Mozammel Chowdhury and Rafiqul Islam and Junbin Gao},
        title={Fast and Robust Biometric Authentication Scheme Using Human Ear},
        proceedings={Security and Privacy in Communication Networks. SecureComm 2017 International Workshops, ATCS and SePrIoT, Niagara Falls, ON, Canada, October 22--25, 2017, Proceedings},
        proceedings_a={SECURECOMM \& ATCS \& SEPRIOT},
        year={2018},
        month={4},
        keywords={Biometric authentication Access control Ear recognition},
        doi={10.1007/978-3-319-78816-6_11}
    }
    
  • Mozammel Chowdhury
    Rafiqul Islam
    Junbin Gao
    Year: 2018
    Fast and Robust Biometric Authentication Scheme Using Human Ear
    SECURECOMM & ATCS & SEPRIOT
    Springer
    DOI: 10.1007/978-3-319-78816-6_11
Mozammel Chowdhury1,*, Rafiqul Islam1,*, Junbin Gao2,*
  • 1: Charles Sturt University
  • 2: The University of Sydney Business School
*Contact email: mochowdhury@csu.edu.au, mislam@csu.edu.au, junbin.gao@sydney.edu.au

Abstract

Biometric authentication using human ear is a recent trend in security applications including access control, user recognition, surveillance, forensic, and border security systems. This paper aims to propose a fast and robust authentication scheme using ear biometric. In this work, a fast technique based on the AdaBoost algorithm is used to detect the ear of the user from profile images. An efficient stereo matching algorithm is used to match the user’s ear data (probe) to the previously enrolled (stored) ear data in a gallery database for verification and recognition. Correspondences are established between extracted features of the probe and gallery image sequences. The performance of the recognition approach is evaluated on different standard ear datasets and compared with other techniques. Experimental results suggest the superiority of the proposed approach over other popular techniques reported in this work.