Ear Based Personal Identification Approach Forensic Science Tasks
Rerkchai Fooprateepsiri* and Werasak Kurutach*** Author for corresponding; e-mail address: * rerkchai@mut.ac.th, ** werasak@mut.ac.th
Volume: Vol.38 No.2 (APRIL 2011)
Research Article
DOI:
Received: 22 September 2010, Revised: -, Accepted: 17 July 2011, Published: -
Citation: Fooprateepsiri R. and Kurutach W., Ear Based Personal Identification Approach Forensic Science Tasks, Chiang Mai Journal of Science, 2011; 38(2): 166-175.
Abstract
Ear has been suggested by the researchers that the shape and features of ear are unique for each person and invariant with age. This paper proposes a robust method for ear based personal identification under environment of variant scaling, rotation and image reflection. Techniques introduced in this work are composed of two parts. The first one is the detection of ear features by using the concepts of multi-resolution Trace transform and Fourier transform. Then, in the second part, modified Hausdorff distance is employed to measure and determine of similarity between the models and tested images. Finally, our method is evaluated with experiments on images from the CMU PIE database. The extensive experimental results show that the average of accuracy rate of ear recognition with variant scaling, rotation and image reflection is higher than 97 %. Keywords: biometric, forensic Science, trace transform, hausdorff distance.