Some forensic aspects of ballistic imaging.

• Author: Cork, Daniel L.

ABSTRACT

Analysis of ballistics evidence (spent cartridge casings and bullets) has been a staple of forensic criminal investigation for almost a century. Computer-assisted databases of images of ballistics evidence have been used since the mid-1980s to help search for potential matches between pieces of evidence. In this article, we draw on the 2008 National Research Council Report (1) Ballistic Imaging to assess the state of ballistic imaging technology. In particular, we discuss the feasibility of creating a national reference ballistic imaging database (RBID) from test-fires of all newly manufactured or imported firearms. A national RBID might aid in using crime scene ballistic evidence to generate investigative leads to a crime gun's point of sale. We conclude that a national RBID is not feasible at this time, primarily because existing imaging methodologies have insufficient discriminatory power. We also examine the emerging technology of microstamping for forensic identification purposes: etching a known identifier on firearm or ammunition parts so that they can be directly read and recovered from crime scene evidence. Microstamping could provide a stronger basis for identification based on ballistic evidence than the status quo, but substantial further research is needed to thoroughly assess its practical viability.

Abstract Introduction I. Firearms Identification A. Toolmarks on Cartridge Casings B. Toolmarks on Bullets C. Uniqueness and Identification Issues II. Computer Images and Assessment of Databases A. Two-Dimensional Technology B. Three-Dimensional Technology C. Signature Analysis D. Scoring and Ranking E. Analysis III. Technical Feasibility of a National Reference Ballistic Image Database A. Performance Studies of IBIS Two-Dimensional Technology B. Panel's Assessment 1. De Kinder Data 2. NBIDE Data 3. NIST Study Overlap Metric C. Is a National RBID Feasible? IV. Microstamping: An Alternative Technology for Tracing to Point of Sale A. What is Microstamping? B. Origins of "Tagging" C. Legislation on Microstamping D. Microstamping of Firearm Parts E. Assessment of the Microstamping Option Conclusion INTRODUCTION

For much of the twentieth century, the forensic science of firearms identification was an intensively individualized activity. A firearms examiner inspected ballistics evidence (spent cartridge cases and bullets) under a comparison microscope, formed a mental pattern of identifying marks and features, and tried to match that pattern against other exhibits. Establishing connections between different cases depended on the memory recall of the firearms examiner or being able to recognize features from photographs in open case files or postings on bulletin boards. (2) Hence, searching through large amounts of ballistic evidence and verifying a match was a labor-intensive and time-consuming task.

Circumstances started to change rapidly in the late 1980s and 1990s as advances in compiling and searching computerized image databases were applied to forensic evidence analysis. The advent of the Federal Bureau of Investigation DRUGFIRE system (for cartridge cases) and the Bureau of Alcohol, Tobacco, Firearms, and Explosives (ATF)-funded BULLETPROOF system (for bullets) made the first significant breakthroughs, permitting individual law enforcement agencies to begin searching new ballistics evidence against large volumes of already-captured images and suggesting possible "hits." (3) The late-1990s formation of the National Integrated Ballistic Information Network (NIBIN), under ATF, broke down the previously formidable geographic barrier by linking the image databases across multiple agencies and permitting searches within regions of the country. (4)

These advances in ballistic imaging technology allow images of bullets or casings to be cataloged, scored, and ranked. A firearms examiner can compare highly-ranked pairs of images on the screen, much as a radiologist might read a digital mammogram. If the matches are promising, the actual physical evidence can be retrieved for direct examination by a firearms examiner for confirmation. (5) The emergence of ballistic imaging also led to the notion of "cold hits"--database searches that could suggest linkages and connections between multiple crimes, without prior knowledge about such linkages--and fueled speculation as to exactly how far the technology could go. (6)

A particularly tantalizing prospect that gained some currency was the creation of a national reference ballistic image database (RBID) that would include images from test fires of all newly manufactured or imported firearms in the United States. (7) At a conceptual level, such a national RBID had the promise of early investigative leads: a cartridge casing recovered from a new crime scene could be entered and searched against the national RBID, generating an investigative lead to at least the initial point of sale of the gun used to fire the round, all without having to actually recover and obtain the gun.

There are several existing ballistic image database resources, but the hypothesized national RBID differed from them in several ways. The NIBIN system and database is not a reference database as it is legally restricted to include only exhibits from firearms used in crime or in law enforcement custody. Other existing RBIDs in the United States are not national in scope; the states of Maryland and New York instituted RBIDs, but those only covered firearms manufactured or sold within those states. (8)

In 2004, the National Institute of Justice (NIJ) requested the National Academies to convene a panel to "Assess the Feasibility, Accuracy, and Technical Capability of a National Ballistics Database" and study the issues surrounding the implementation of a national RBID. The findings were published in the Report Ballistic Imaging, (9) which predates Strengthening Forensic Science in the United States: A Path Forward. (10) The panel, on which the authors of the current Article served in different capacities, included experts from a wide range of scientific disciplines bearing on the application of imaging techniques and computerized search to assist firearms identification. (11) The disciplines covered include statistics, mechanical engineering, materials science and metallurgy, computer science, and manufacturing design. The panel's deliberations focused on the assessment of three basic policy options:

* Maintain the existing NIBIN system as-is, with its restriction to crime scene evidence;

* Enhance the NIBIN system in operational or technical ways to facilitate better matching of evidence images; or

* Develop a national RBID including images from test fires of all newly manufactured or imported firearms. (12)

One major technical enhancement identified early in the study process was using three-dimensional surface measurements as "images" of ballistics evidence (and their associated toolmarks) instead of two-dimensional photography. To that end, the panel supervised a program of experimental work by the National Institute of Standards and Technology (NIST), funded under a separate NIJ grant. (13)

The panel also chose to investigate the emerging technology of micro-stamping: etching microscopic identifiers onto ammunition components or firearm parts so as to impose known markings onto ballistics evidence. This approach could also provide early investigative leads without requiring recovery of the crime gun using an alternative technology. (14)

This Article summarizes some of the policy themes in Ballistic Imaging, as an argument for the kinds of research on conceptual foundations in forensic science envisioned in Strengthening Forensic Science. We discuss the limitations and potential of technology as they apply to a particular forensic science application. In terms of the construction of a national RBID and the implementation of firearms microstamping, the panel concluded that both concepts were premature and needed more and substantial research. (15)

In Part I, we outline some basic issues in firearms identification, including the types of toolmarks and the problem of uniqueness. Part II describes the existing ballistic imaging hardware and software platform reviewed by the panel, the Integrated Ballistics Identification System (IBIS) and the means by which it compares and scores potential matches. Part III summarizes the original experimentation results and extant performance studies that contributed to the panel's central finding that a national RBID is inadvisable at this time. Part IV covers the state of the alternative technology of microstamping at the time of the panel's study, and we conclude with some general themes. Before we proceed further, some caveats are in order:

* An assessment of the admissibility of forensic firearms evidence in legal proceedings was not part of our charge and was beyond the expertise of our panel. This is clearly an important topic that is covered in other articles in this issue.

* The current use of ballistic image databases requires comparison of the physical exhibits by a firearms examiner to certify a "hit" or match--presumably to ensure that someone can testify to the "hit" in legal proceedings. (16) Accordingly, we did not consider the issue of whether human firearms examiners might be replaced by a mechanical routine.

* Our study was "neither a verdict on the uniqueness of firearms-related toolmarks generally nor an assessment of the validity of firearms identification as a discipline." (17) Rather, we were asked to study "the uniqueness of ballistic images"--the uniqueness and reproducibility of the toolmarks on ballistics evidence as they are recorded or measured by various technologies (e.g., photography or surface measurement). (18)

1. FIREARMS IDENTIFICATION

   Toolmarks are created when a hard object (generally, a tool) impacts a relatively softer object. In the case of modern firearms and ammunition, those marks are generated in...