This paper describes experiments with developing a semi-automated method to compare 3D dental models taken from candidate humans and bite mark impression images left in the scene of the crime.
Forensic dentistry generally addresses the problem of identifying individuals based on the properties of teeth or identifying individuals based on bite mark impressions. It is legally relevant to accurately and reliably match a bite mark impression to place a criminal at the scene of a crime. Therefore, a system which minimizes human interaction to conduct the comparison would be beneficial to ensure accuracy and reduce human bias. In the current study, once the contours from the bite mark image and the 3-dimensional dental model are captured, the ideal alignment is calculated by finding the transformation which minimizes a distance measure. The best match is then identified by performing this comparison to a set of candidate dental models. The results are compared to identification results by human forensic odontology experts. (Publisher abstract provided)