Bloodstain pattern analysis as optimisation problem

Abstract

Bloodstain pattern analysis (BPA) is an approach to support forensic investigators in reconstructing the dynamics of bloody crimes. This forensic technique has been successfully applied in solving heinous and complex murder cases around the world and, recently, computer-based BPA approaches have been designed to better support investigators both in terms of speed and quality of analysis. However, despite its widespread use, current automatic techniques for BPA try to define some algorithmic steps to replicate a sequence of subjective investigators' tasks without relying on any mathematical formalism to compute an objective reconstruction of the crime. The lack of an objective mathematical foundation is a critical issue in a scenario where the quality of evidences can strongly affect a court trial and the life of people involved in that trial. This paper introduces the very first formal representation of BPA by means of an optimisation problem, on which to base the next generation of crime reconstruction techniques. As an example of the benefits provided by the proposed formal representation of BPA, a case study based on a genetic algorithm shows how the BPA optimisation problem can support investigators in performing a fast, precise, automatic and objective analysis.

Keywords: Bloodstain pattern analysis; Forensic sciences; Genetic algorithms; Optimisation.