Use of pattern recognition and neural networks for non-metric sex diagnosis from lateral shape of calvarium: an innovative model for computer-aided diagnosis in forensic and physical anthropology

Abstract

Sex determination on skeletal remains is one of the most important diagnosis in forensic cases and in demographic studies on ancient populations. Our purpose is to realize an automatic operator-independent method to determine the sex from the bone shape and to test an intelligent, automatic pattern recognition system in an anthropological domain. Our multiple-classifier system is based exclusively on the morphological variants of a curve that represents the sagittal profile of the calvarium, modeled via artificial neural networks, and yields an accuracy higher than 80 %. The application of this system to other bone profiles is expected to further improve the sensibility of the methodology.

Appendix: The algorithm for probability density estimation

Let x ₁, …,x _n be a collection of n CT-scans, thought of as d-dimensional random vectors and assumed to be independently and identically drawn from an unknown pdf p(.). Also, let φ(.) be a proper kernel function (e.g., a standard Gaussian), and let the corresponding bandwidth h ₁ be any positive real number (to be fixed empirically) [2]. An unbiased estimate ${\tilde p}(.)$ of p(.) via MLP is proposed, according to the following unsupervised algorithm (expressed in pseudo-code):

where $\frac {1}{n-1} {\sum }_{\textbf {x} \in \mathcal {T}_{i}} \frac {1}{V_{n-1}} \varphi \left (\frac {\textbf {x}_{i} - \textbf {x}}{h_{n-1}}\right )$ is the Parzen kernel expansion of p(.) over the n−1 feature vectors in $\mathcal {T}_{i}$, V _n−1 being the corresponding volume of the window function [6].