QUESTIONED DOCUMENT EXAMINATION AND FIREARMS

Questioned Document Examination AND Firearms and Tool mark Examination

Questioned Document Examination and Firearms and Tool mark Examination

Questioned Document Examination

The fields of questioned document examination would benefit greatly from a sharing of a common knowledge base. It can offer technologically advanced methods to facilitate, enhance and validate the work of a document examiner. On the other hand, the accumulated knowledge gained by document examiners over more than 100 years is extremely valuable for the creation of new engineering solutions for the writer identification problem (Conway, 2005).

Furthermore, it is also clear that in forensic handwriting analysis, several important factors can influence the behavioral patterns which underlie the handwriting process - for example, the occupational characteristics of the writer, national or ethnic origin, even the task context in which the writing is executed - and a better understanding of such factors and their influence could be important and valuable in seeking greater integration across the boundaries of the forensic/ engineering communities (Hilton, 2002).

This paper represents a step in this direction and, in particular, seeks to explore the nature, variability and discriminatory power of a range of commonly extracted features of handwriting in relation to some of the factors which characterize the writer. The forensic document analysis community currently does not employ most of the dynamic features which have been shown to be very successful in engineering applications of handwriting analysis. We will explore some of the potential benefits associated with dynamic analysis and, in particular, how these features vary when writers perform different types of writing tasks (Harrison, 2006).

However, it must also be recognized that it is precisely these dynamic features which are not generally available in most forensic analysis tasks. This, then, points to the possible benefits of developing a strategy based on the better exploitation and effective deployment of static handwriting features, or the increasing refinement of techniques to predict 'pseudo-dynamic' features from static measurements (Conway, 2005).

Similarly, automatic handwriting analysis specialists can benefit from a better understanding of how the constraints imposed by different task environments and objectives can influence the way in which handwriting tasks are executed, particularly, in relation to the effect these have on the analytical power of different extractable features (Hilton, 2002). Thus, in the study reported here, we wish to investigate the stability and discriminatory capability of a number of writing features commonly extracted from (generally small-scale) writing samples. In particular, we wish to analyze the difference in the analytical power of features across a range of different writing tasks, across different population demographics and across subjects, both as they produce their own handwriting and while imitating (as perhaps a criminal forger might attempt) handwriting samples from other writers (Harrison, 2006).

The paper is organised in three further sections, describing the compilation of a database of widely varying handwriting samples, presentation of an analysis of the data and, finally, a discussion of some conclusions which can be drawn from the ...