2021
Xiaoyu Alan Zheng
Miami, FL
The last decade has seen exciting progress in the development of measurement instruments, algorithms, data, and methods to facilitate objective analysis of toolmark comporisons. The primary goal is to provide firearms and tool mark examiners the ability to support their testimonies with objective similarity values and statistically-sound quantitative expressions for the weight of the evidence. The National INstitue of Standards and Technology (NIST), Federal Bureau of Investigation (FBI), and the Netherlands Forensics Institute (NF) have been collaborating to the develope the REference Population Database of FIrearm Toolmarks (RPDFT). RPDFT is a system that consists of a reference database of firearms toolmarks (impressed and striated), data processing modules, quantitative similarity metrics, statistical weight of evidence calculation protocols, and a user interface. In the past three years, a reference database infratructure have been developed to index test fires generated by the FBI according to thier class characteristics. The reference database consists of ground truth KM and KNM comparisons and is designed to be filtered according to matching/relevant class characteristics of the evidence being analyzed. NIST and NFI’s objective toolmark analysis algorithms and statistical protocols have also been integrated into the database. These allow for quantitative one to one comparisons of firearm toolmarks which are then used in conjunction with the relevant reference populations to calculate its statistical weight of evidence. The presentation will describe the reference database desing, analysis moduls, proposed workflow/implementation, statistical analysis, milestones and future plans of RPDFT.