Advanced Search

Home > Journals > Minerva Medicolegale > Past Issues > Minerva Medicolegale 2009 December;129(4) > Minerva Medicolegale 2009 December;129(4):191-217



A Journal on Forensic Medicine

Frequency: Quarterly

ISSN 0026-4849

Online ISSN 1827-1677


Minerva Medicolegale 2009 December;129(4):191-217


Applying statistical pattern comparison methods to tool marks

Petraco N. D. K. 1, Petraco N. 2

1 John Jay College of Criminal Justice, Chemistry PhD. Program, CUNY Graduate Center
2 NYPD Criminologist, Retired, Criminal Justice Ph D. program, CUNY Graduate Center

In February 2009 the National Academy of Sciences of the United States published a report entitled “Strengthening Forensic Science in the United States: A Path Forward” which addresses ways to build on the current state of knowledge and practices in the forensic sciences. The report notes research studies must be performed to “…to understand the reliability and repeatability…” of comparison methods commonly used in forensic science. This paper is a review of many available, statistically based methods and practices which can be applied in short order to tool mark impression patterns. The power of these methods is that they can assign objective, quantitative measures to the words “comparative examination” and “sufficient agreement”. Their outputs are called decision rules and can be found with two different approaches, classical methods and postclassical methods. The classical methods are based on knowledge of probability distribution functions and include popular “Bayesian” methods. The second general approach, postclassical methods, do not rely directly on probability distribution functions. Rather try to “learn” pattern identities directly from data. Both approaches have strengths and weaknesses, but all methods described are used and tested every day in industry and academia.

language: Italian


top of page