NCJ Number
227696
Journal
Journal of Forensic Sciences Volume: 54 Issue: 4 Dated: July 2009 Pages: 761-772
Date Published
July 2009
Length
12 pages
Annotation
This paper proposes a new digital method for the objective comparison of frontal sinuses for identification.
Abstract
Correlation coefficients with the proposed method demonstrated inter-rater and test-retest reliability. Further refinement of the reference distributions and more rigorous testing of error rates should make this technique applicable to casework. Use of the dimensions of the frontal sinuses as a means of identifying individuals requires an objective method of comparison in order to meet Daubert standards for evidence. Christensen's application of Elliptical Fourier Analysis and Likelihood Ratios seems to be a viable solution for this problem. The proposed method used in the current study draws on this work and simplifies its application. Variation between pairs of digitized sinus tracings was quantified by summing the difference between corresponding measurements taken from a fixed origin to the outer edge of the sinus outlines, using Adobe Photoshop CS2. Same-skull and different-skull pairs were used to develop reference distributions from which the probability of unknown pairs coming from the same or a different individual was estimated. Three adult samples were used in this study, archeological, clinical, and contemporary. The archeological sample consisted of a radiographic collection of 46 adult skills obtained from the Laurentian University Forensic Osteology Laboratory. The clinical sample was obtained from the Sudbury Regional Hospital and consisted of 63 documented P-A sinus radiographs. This sample consisted of adults of mixed ancestry typical of the Greater Sudbury area. The contemporary sample consisted of 16 adult skulls of a contemporary origin, part of the Laurentian University Department of Forensic Science human osteology teaching collection. This paper describes the tracing procedure, image analysis, calculation of total difference, intraobserver and interobserver error, frequency distributions, blind analysis, and empirical testing. 4 tables, 11 figures, and 41 references