U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Image Registration-Based Approach to Ranking Dental X-Ray Images for Human Forensic Identification

NCJ Number
225685
Journal
Canadian Society of Forensic Science Journal Volume: 41 Issue: 3 Dated: September 2008 Pages: 125-134
Author(s)
Maja Omanovic; Jeff J. Orchard
Date Published
September 2008
Length
10 pages
Annotation
This paper presents a computer-aided framework for matching dental radiographs based on a sum of squared differences (SSD) cost criterion.
Abstract
Results show that matching dental records using the SSD cost function is a viable method to aid in human forensic identification. The approach presented is versatile and can easily be incorporated into the existing systems whether in a serial or parallel mode. The two methods can be run separately in parallel and their identification search results combined or, one approach can be used to pre-filter the database and the other to refine the results. Noted, however, is that in order to fulfill the needs of future forensic identification systems, a move away from text-based approaches is desirable. The experiments were executed on a typical desktop computer, a Pentium 4 Processor 2.8 Ghz machine with 1G of RAM. The comparison between the two images takes approximately 0.5 seconds. In 93 out of 150 cases (62 percent of the trials), the correct match was in the top 1 percent. In 90 percent of the identification trials, the method ranked the correct match in the top 10 percent. Data were collected using MATLAB (Math Works Inc., Natick, MA) and tested on a database containing 571 dental radiographs belonging to 41 distinct individuals. Figures and references