A variant approach for human forensic identification using dental radiographs with skeleton and contour

Abstract

Forensic odontology is a branch under biometrics in which victim identification is performed with dental radiographs. It is made by matching ante–mortem and post–mortem dental records. In this paper, a skeleton–based shape matching algorithm for dental bitewing and periapical radiographic images is proposed. To improve the illumination evenness of the dental radiographs, a combination of Butterworth and homomorphic filters is used in the pre–processing stage. Spline function is used for teeth isolation. Missing tooth in dental radiographs is obtained by using bifurcation technique. An active contour model with Selective Binary and Gaussian Filtering Regularised level set method is used for contour extraction. Shape is also extracted by using skeleton. Finally, shape matching is achieved by both contour–based and skeleton–based approaches. The experimental results are obtained on a database of 100 ante–mortem and 50 post–mortem images. This algorithm provides a hit rate of 0.77 for both the contour and the skeleton approaches. The retrieval rate is comparatively better for contour than skeleton for other rankings.