CT methodologies and mathematical algorithms for characterising items of forensic relevance

Project summary

Introduction

Forensic science is in a state of crisis.  Many long established techniques accepted by the courts have been shown to be deficient in underpinning scientific robustness. This has created a crisis of confidence in the validity of such analysis and their consequent presentation as evidence in criminal cases. The relatively recent application of technologies from other fields (eg CT from medicine) to forensic examinations offers an opportunity to redress these scientific deficiencies. We will develop appropriate scientific and interpretative frameworks to underpin the use of CT imaging for the analysis of marks made on a range of surfaces by implements such as knives. We will produce data sets of known origin from a range of analytical instruments and develop mathematical models which can be used to differentiate the implements used to create the marks. Aligned to this is the development of predictive tools identifying potential implements given the marks observed.

Project aims

The purpose of this application is to collaboratively develop a new forensic capability in Scotland which will be ambitious, ground breaking and unique. This new capability will initially address one specific aspect of forensic identification evidence, that of tool mark identification, for creating a methodological template with eventually far wider applications. By combining our demonstrated forensic research excellence and understanding of the requirements of criminal justice end-­‐users with the distinct technical expertise and capabilities of X-­‐ray computed tomography (CT) and of image analysis, we will develop and assess the application of CT devices to interpretation of damage patterns caused on various substrates (for example wood, fabric, bone) using a range of implements (knives and saws).

The project provides an opportunity for the development of a research focus  around  the  specific  areas  of  forensic science where modern technologies can be applied to address what are often complex pattern recognition and interpretation problems. This involves a multi stage process. Firstly, data of known provenance data need to be captured in a methodologically sound and scientifically robust way in order to generate a ground truth database. Secondly, mathematical algorithms need to be developed, evaluated and validated to exploit the developed database as an interpretative tool for aiding in the potential classification of unknown samples.