Title: MRI-based knee image for personal identification
Authors: Lior Shamir
Addresses: Department of Computer Science, Lawrence Technological University, 21000 W Ten Mile Rd., Southfield, MI 48075, USA
Abstract: While external body parts such as the face, fingerprints, or retina are often used for biometric identification, it can also be reasonably assumed that internal organs imaged with biomedical imaging devices can also allow biometric identification. Here we studied the use of MRI images for the purpose of biometric identification, and show that the accuracy of person identification using knee MRIs is significantly higher than random. Knee MRI images of 2,686 different patients were used in the experiment, and analysed using the wndchrm image classification scheme. Experimental results show that the rank-10 identification accuracy using the MRI knee images is ∼93% for a dataset of 100 individuals, and ∼45% for the entire dataset of 2,686 persons. Since MRI is used for the purpose of imaging internal parts of the body, this approach of biometric identification can potentially offer high resistance to deception.
Keywords: MRI images; magnetic resonance imaging; image analysis; internal biometrics; knee images; medical imaging; personal identification; biometric identification.
International Journal of Biometrics, 2013 Vol.5 No.2, pp.113 - 125
Received: 06 Sep 2011
Accepted: 05 Jun 2012
Published online: 28 Feb 2014 *