How Statistical Pattern Comparison Methods Can Be Applied to Toolmarks

Over the course of the last two decades, DNA profiling has grown to be one of the most widely known and applied techniques for the identification of biological samples in forensic science. Arguably, the “fame” of DNA profiling is responsible for the courts’ current interest in the raising of standards for scientific examination of all forms of physical evidence (e.g., toolmarks, soils, fingerprints, gunshot residue, etc.). It is undoubtedly the clear applicability of simple statistical methods to DNA profiling techniques that lies at the core of its success. Unfortunately, there are no standard methods or protocols for the application of probability and statistics to the analysis and comparison of toolmarks. In order to begin to address this problem, here we propose and describe several methods for objective, numerical computational pattern matching that can be applied to toolmarks.* Indeed, Moran remarks that “the actual mechanics of the toolmark identification process involves the use of our cognitive ability to first recognize identification through pattern matching.”1 The theory and algorithms of statistical pattern recognition give methods for what Moran calls “the quantitative difference between an identification and non-identification.”1 Industries, particularly those that employ some type of sorting into categories, face similar problems as those faced in the examination of physical evidence in forensic science. They have been applying statistical pattern recognition techniques for the automation of assorted classification tasks for approximately the last half century.2–6 The examination of similarities and dissimilarities between pieces of physical evidence is based on comparison of the evidence’s physical features. The Association of Firearms and Toolmark Examiners (AFTE; www.afte. org) is the de facto group that has set the standards for firearms and toolmark examination. The AFTE theory of identification as it relates to toolmarks states (boldface added for emphasis):1,7