Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 9
Ballistic Imaging PART I Context for Ballistic Imaging Analysis
OCR for page 10
Ballistic Imaging This page intentionally left blank.
OCR for page 11
Ballistic Imaging 1 Introduction For decades, direct comparison of bullet and cartridge case evidence has been used to link crime incidents to other crime investigations to link specific pieces of evidence to each other and to particular weapons. Since the late 1980s, emerging technology has allowed such links to be drawn between computerized images of bullets and cartridge case evidence. The development of this technology has led to speculation about its potential to generate critical investigative leads to possibly related incidents both at the local level and across broad geographic areas. A specific question that has been raised concerns the utility of a national reference ballistic image database (RBID), which would include images from test-fired rounds of most (if not all) new and imported firearms. In concept, a national RBID would permit bullet or cartridge case evidence recovered at crime scenes to be easily and rapidly linked to a firearm’s point of sale—information that is currently available only if the gun itself is recovered at the crime scene and is put through a full tracing process (see Chapter 9). The concept of a national RBID differs from existing systems in two important ways. First, a national database of ballistic image evidence already exists, but it is not a reference database because it does not collect test firings from new weapons. In 1997, the Bureau of Alcohol, Tobacco, Firearms, and Explosives (ATF) formed the National Integrated Ballistic Information Network (NIBIN); as of 2005, NIBIN connects 230 law enforcement agencies, which contribute to a database of images of bullet and cartridge case evidence recovered from (or test-fired from weapons linked to) crime scenes. The NIBIN program equips agencies with Integrated Ballistics Identification System (IBIS) equipment, developed and
OCR for page 12
Ballistic Imaging manufactured by its contractor, Forensic Technology WAI, Inc. (FTI), of Montréal, Canada (see Box 4-1 in Chapter 4 for an important note on the use of the term “IBIS”). The IBIS platform acquires greyscale photographs of bullet or cartridge case evidence, scoring and ranking pairs of exhibits by deriving a mathematical signature from images. The scope of NIBIN is limited by legislative language, prohibiting it from including noncrime gun evidence in the database. Second, RBIDs do currently exist but not at the national level. Two states—Maryland and New York—established RBID systems for new handguns sold in those states in 2000 and 2001, respectively. Both states use the same IBIS platform for acquiring images, but are barred from directly networking their RBID data with crime-scene-based NIBIN data. State legislation directed the California Department of Justice to study the feasibility of establishing an RBID in that state; its assessment in 2001 was that such a database was not feasible, but suggested that further study be conducted at the national level. In the wake of the October 2002 sniper shootings in the Washington, D.C., area, legislative proposals to create RBIDs were advanced or discussed in Connecticut, New Jersey, and Massachusetts (Butterfield, 2002), as well as Missouri (George, 2004a).1 As of 2002, a national-level RBID was said to be under discussion in Belgium, and “a similar debate is going on in a number of member states of the European community” (De Kinder, 2002a:198). Recent U.S. Congresses have seen bills introduced that would create a national RBID, though none of the bills have advanced past referral to the appropriate committees.2 In 1 Both chambers of the New Jersey state legislature have, at different times, passed bills requiring some form of ballistic imaging, but not the same bill. In May 2000, the Senate passed S. 2048 on a 37–0 vote; the bill prohibited sales of handguns unless a “ballistics identifier” was obtained from the gun and put in a “qualified database.” A “ballistics identifier” was defined as “a digitized or electronic image of a bullet and shell casing … clearly showing the distinctive firing pin, ejection, extraction and land marks for that particular handgun.” In November 2002, the Assembly passed A. 438 on a 48–18–10 vote; initial bill text made submissions to an image database voluntary by handgun owners, but the passed bill had been amended to make submission of identifiers mandatory for all sold handguns. The bill was not acted upon by the Senate. In the 2004–2005 session, proposed bills would have required firearms repair shops to obtain ballistics identifiers for handguns or rifles before returning them to their owners; as of September 2006, no similar bills had been introduced in either chamber. In Massachusetts, a Boston police official lauded the idea as a “great law enforcement tool,” pointing to a case that had been solved using NIBIN (linking the same .22 caliber Ruger pistol to shootings of seven people in four cities,” as one where an earlier investigative lead to the gun’s purchaser would have been useful (Butterfield, 2002). 2 See, e.g., in the 108th Congress, the Technological Resource to Assist Criminal Enforcement (TRACE) Act (S. 469/H.R. 776) and the So No Innocent Person Ever Repeats (SNIPER) the Sniper Tragedy Act of 2003 (S. 1983), the latter of which incorporated the former in its entirety, as well as the Bullet Tracing Act to Reduce Gun Violence Act (H.R. 24). In the 107th Congress, see the Ballistics, Law Assistance, and Safety Technology (BLAST) Act (H.R. 5663)
OCR for page 13
Ballistic Imaging part, the bills did not progress because they have been caught up in the nation’s ongoing gun control policy debate: Proponents see such laws as essential to reducing gun violence; opponents see them as a first step toward a national gun registry and a perceived violation of their Second Amendment right to bear arms. In both the 107th and 108th Congresses, bills were introduced in each chamber to require that the National Academies conduct a study of the state of ballistic imaging technology.3 Independent of that legislation, but with a similar charge, the National Institute of Justice (NIJ) of the U.S. Department of Justice requested that the National Academies execute such a study. 1–A COMMITTEE CHARGE In 2004, as requested by NIJ, the National Academies appointed the Committee to Assess the Feasibility, Accuracy, and Technical Capability of a National Ballistics Database. The committee was asked to: assess the feasibility, accuracy and reliability, and technical capability of developing and using a national ballistics database as an aid to criminal investigations. To accomplish this, the [committee] will Assess the technical feasibility, through analysis of the uniqueness of ballistic images, the ability of imaging systems to capture unique characteristics and to parameterize them, the algorithmic and computational challenges of an imaging database, the reproducibility of ballistic impressions and the ability of imaging systems to extract reproducible information from ballistic impressions. Assess the statistical probabilities that ballistics evidence presented would lead to a match with images captured in a database, whether and how the base rate can be estimated for those crimes that present bullet or casing evidence that do in fact come from a gun that produced a database entry, and the probabilities and consequences of false positives and false negatives. Assess the operational utility of ballistics evidence in criminal investigations—that is the extent to which it is used or can be used to identify crime guns and suspects and to solve specific crimes. Assess the sources of error in ballistics database matching (from examination, digitization, computer matching, chain of custody and documentation of tests, and expert confirmation), how they may be quantified, and how these errors interact. and the Bullet Tracing Act to Reduce Gun Violence Act (H.R. 422). Earlier versions of the Bullet Tracing and BLAST Acts were also introduced in the 106th Congress. 3 See H.R. 3491 and S. 2581 in the 107th Congress and H.R. 2436 and S. 980 in the 108th Congress. These bills also failed to advance beyond referral to subcommittees.
OCR for page 14
Ballistic Imaging The charge continues: The committee’s work will provide scientific and technical knowledge to inform the government’s deliberations on three policy options with regard to ballistics databases: Maintain the National Integrated Ballistic Information Network (NIBIN) on ballistics recovered from crime scenes. It is operated by the Bureau of Alcohol, Tobacco, and Firearms.4 Enhance the NIBIN system so that it can be used to match crime scene evidence with the gun used. Establish a national ballistics database of images from bullets fired from all, or nearly all, newly manufactured or imported guns for the purpose of matching ballistics from a crime scene to a gun and information on its initial owner. We note that the committee was specifically tasked by NIJ to consider these policy options on the basis of the scientific evidence of system performance and not to include or exclude options based solely on their cost. That is, assessing the cost-effectiveness of ballistic imaging and related techniques is not a dimension of our charge. The wording of the charge raises a few questions. In several instances—as in the formal name of the committee—the term “ballistics” is used in an imprecise manner; see Box 1-1. The charge also provides no specific direction on how the existing NIBIN system might be “enhanced” in its second policy option. The third policy option is also somewhat imprecise in representing basic assumptions about the nature and intent of a national RBID; see Box 1-2 (see also Section 9–B in Chapter 9). At the outset, it suffices to say that reasonable proposals for a national RBID would most likely focus exclusively on images of cartridge casings (not bullets, as described in the charge) due to the longer time necessary to recover and process test-fired bullets, and—at least at the outset—would likely be further restricted to samples from handguns and small arms. The charge also suggests that the intent of a national RBID is to provide “information on [a gun’s] initial owner.” That is certainly the goal of criminal investigations that would make use of an RBID, but the RBID search itself would be intended to provide an investigative lead to a point of sale, one step removed from information on the initial owner. As with the current gun tracing system, additional investigative work based on the point of sale would be needed to determine a gun’s initial ownership; as discussed in Chapter 9, the content of a national RBID does not necessarily involve entering purchaser-specific data. 4 In January 2003, the agency was renamed the Bureau of Alcohol, Tobacco, Firearms, and Explosives, though it commonly retains the acronym ATF.
OCR for page 15
Ballistic Imaging BOX 1-1 “Ballistics” Terminology Ballistics, literally, is the study of the dynamics of projectiles in flight. It is not equivalent to “firearms identification,” though in common usage, as in the title of this committee, the word has been interpreted that way. Calvin Goddard—considered the father of modern firearms identification—was chagrined at his own role in initiating this use of language. He titled his landmark 1925 paper on the use of the comparison microscope “Forensic Ballistics,” a name selected “after long and prayerful consideration, and in an effort to employ terms that would be concise and at the same time meaningful;” it was, however, “a title that has plagued me ever since” (Goddard, 1999:233): Forensic was good enough, since it means that which has to do with public disputation, and was what I meant to say. “Ballistics” was bad, very bad, since ballistics strictly used applies solely to projectiles in motion, and the forces that influence that motion. Thus far, I never made an attempt to identify a projectile in motion, and if I ever have to, it will be too soon, so far as I am concerned. However, the man in the street found ballistics [an] interesting word, and seized upon it avidly, at the same time discarding the “forensic” which, when used jointly with ballistics, partly takes the curse off the latter. Likewise, Hatcher (1935:20) rued the way “the word ‘ballistics’ has in the past several years become associated in the public mind with the science of Firearms Identification.… I realize fully that usage makes language, and that the recent rather extensive mis-use of the word Ballistics in this way may be a valid excuse for continuing the practice; but still it seems to me that the use of the word to describe the Science of Firearms Identification is somewhat undesirable in any case, as being loose English.” Forensic scientists distinguish between four types of “ballistics” (Rinker, 2004): Internal ballistics refer to the forces—pressure, ignition, and so forth—that operate on the bullet while still inside the firearm. External ballistics, closest to the literal definition of ballistics, describes the flight of a bullet between the firearm muzzle and its impact at target. Terminal ballistics describe the mechanics of impact on both the projectile and the target. Forensic ballistics, in Goddard’s sense, is the analysis of bullet and cartridge case evidence and the use of that evidence to link specimens to each other and to particular weapons. “Ballistics” is convenient shorthand but in this report—save for the committee’s formal name—we try to refrain from the use of the word on its own. Our use of the adjective “ballistic” (as in “ballistic imaging” and “ballistics evidence”)—like any instances of “ballistics” that may still appear in the text—is properly interpreted as referring to “forensic ballistics.”
OCR for page 16
Ballistic Imaging BOX 1-2 Content of a Reference Ballistic Image Database A major part of our committee’s charge is to assess the feasibility of a reference ballistic image database. However, many of the parameters governing the content of such a database are not specified by the charge—for instance, whether cartridge casings or bullets (or both) should be entered into the database, whether one bullet or casing is sufficient or whether multiple exhibits are needed, whether all firearm types should be covered. The bills introduced at the federal level have been worded to be as inclusive as possible. In the 108th Congress, the BLAST Act, SNIPER Act, and TRACE Act (H.R. 5663, S. 1983, and H.R. 776, respectively) bore identical language on the nature of the envisioned database. “A licensed manufacturer or licensed importer” would be required to test fire firearms manufactured or imported by such licensees as specified by the Attorney General by regulation; prepare ballistic images of the fired bullet and cartridge casings from the test fire; make the records available to the Attorney General for entry in a computerized database; and store the fired bullet and cartridge casings in such a manner and for such a period as specified by the Attorney General by regulation. The database envisioned by these bills would require both bullets and casings to be entered; it would also put the responsibility for image acquisition and exhibit archival on the manufacturers or importers. The question of whether image data would be collected for long guns as well as handguns was not directly answered. The existing state reference ballistic image databases in Maryland and New York both made key limiting assumptions, restricting their content to cartridge In structuring our work, we have taken the three policy options as a guide; addressing them necessarily involves addressing the issues suggested in the preceding four substantive points of the charge. Cast in language more consistent with usage in the field, we have interpreted our principal task as providing information on three different federal policy options: Maintain the NIBIN as it presently exists—that is, retaining the restriction that only crime-gun-related evidence be included in the database. Enhance the current NIBIN system in order to increase its effectiveness without expanding its scope to include new or manufactured firearms; such improvements could include changes to the basic imaging standard used by the system (e.g., three-dimensional surface measurement rather
OCR for page 17
Ballistic Imaging casings from handguns only. The enabling law in Maryland created a centrally located “Statewide Shell Casing Data Base,” later called MD-IBIS, and a “Statewide Shell Casing Repository,” both to be administered by the Maryland State Police Crime Laboratory. The shell casings in question were to be “provided by dealers from all handguns sold in the state” and transmitted to the Crime Laboratory (Maryland Code 29.05.02.02). In later years, a bill to expand MD-IBIS coverage to long guns was introduced in the legislature but was not enacted. Similarly, the New York Combined Ballistic Identification System (CoBIS) database was established as the “pistol and revolver ballistic identification databank” (New York General Business Laws, Article 26, Section 396-ff): Any manufacturer that ships, transports or delivers a pistol or revolver to any person in this state shall … include in the container with such pistol or revolver a separate sealed container that encloses: (a) a shell casing of a bullet or projectile discharged from such pistol or revolver; and (b) any additional information that identifies such pistol or revolver and shell casing. The language of the previous federal bills notwithstanding, we generally assume throughout this report that a national reference ballistic image database would be similar to the Maryland and New York models albeit at the larger, national scale. At the minimum, we assume that operational constraints would limit the national reference database to cartridge casings, owing to the time-consuming process of discharging weapons in a water tank or other trap so that expended bullets can be recovered in “pristine” condition. Whether rifles and long guns would be included in such a database is an open question; again, the Maryland example leads us to assume that initial coverage would be focused on handguns (as the major class of guns used in crime). than two-dimensional photography), improvements to database handling, improved procedures for working with the existing hardware and software, and so forth. Establish a national reference ballistic image database, as a counterpart to (and possibly linked to) NIBIN, containing images of ballistic samples from all newly manufactured or imported guns, in order to generate investigative leads to the original point of sale of a firearm. 1–A.1 Experimental Study by National Institute of Standards and Technology In support of the committee’s work, NIJ separately contracted with the Office of Law Enforcement Standards of the National Institute of Standards
OCR for page 18
Ballistic Imaging and Technology (NIST) for experimental work with the direction and advice of the committee, on the feasibility and accuracy of identification using a national ballistic image database. In particular, the NIST work considers the relative advantages of three-dimensional metrology techniques as compared to the current two-dimensional imaging used in NIBIN. NIST’s experimentation in support of this study builds on NIST’s ongoing work, under other contracts, on the development of standard bullets and cartridge casings for the calibration of ballistic imaging systems. The NIST work is summarized in Chapter 8, and the full NIST report has been published separately (Vorburger et al., 2007). 1–A.2 Limitations: What the Committee Study Does Not Do In the balance of this chapter, we provide additional basic context for the committee’s work and give an overview of the structure of the report. However, we believe that it is first important to be clear on certain limitations of our work and our charge. Our task is to assess various policy options related to ballistic imaging; it is possible for this basic charge to be misconstrued or overinterpreted in at least three major ways. First, and most significantly, this study is neither a verdict on the uniqueness of firearms-related toolmarks generally nor an assessment of the validity of firearms identification as a discipline. Our charge is to focus on “the uniqueness of ballistic images”—that is, on the uniqueness and reproducibility of the markings (toolmarks) left on cartridge cases and bullets as they are recorded or measured by various technologies (e.g., photography or surface metrology). The uniqueness of firearms-related toolmarks generally is a much broader question—and a very important one—but it is not one that our committee was constituted to address. At a minimum, assessing the general validity and uniqueness of toolmark evidence would require a much wider range of gun and ammunition selections and firing conditions than was supported in our experimentation through NIST (see Chapter 9). It would also require precise quantification of the myriad sources of variability inherent in the firing of a gun (see Chapter 2). In short, it would be a major undertaking, requiring a sustained program of research over many years, and it is impossible to definitively answer the question of the uniqueness of ballistic toolmarks as a by-product of a more targeted study of the uniqueness of ballistic images. Although a definitive statement on firearms toolmark uniqueness is not within our purview, some discussion of issues related to uniqueness—particularly the sources of variability in generating such toolmarks—are essential to our work. Chapters 2 and 3 of this report are largely dedicated to these matters, covering the sources of variability inherent in firing a
OCR for page 19
Ballistic Imaging gun and the uniqueness and reproducibility of firearms-related toolmarks as judged by firearms examiners using the comparison microscope. From these reviews, some readers may attempt to infer a stance by this committee, for or against the validity of firearms identification generally. From this perspective, some may argue that our narrow focus on the uniqueness of ballistic images amounts to missing the proverbial elephant standing in the room: that is, we should extend any conclusions on the strength or weakness of ballistic image evidence to infer the strength or weakness of ballistic toolmark evidence more globally. We reiterate that no such broader conclusion is intended by this report, which was not developed to support more sweeping statements. Rather, the examination in Chapters 2 and 3 is intended to identify major sources of variability, as well as particularly challenging (or easy) contexts for linking pieces of ballistics evidence, in order to best construct our experimental work with NIST and understand the results of ballistic image database comparisons. Other readers may see a definitive statement on the uniqueness and reproducibility of toolmarks as a first and necessary building block to any further work: without such a statement—if firing processes and resulting toolmarks are completely random—then the basic utility of a ballistic image database (of any sort or scope) to try to suggest connections between pieces of evidence comes into question. We appreciate this argument but conclude that it is possible to speak meaningfully about ballistic image database performance without first fully accepting or concluding the fundamental uniqueness of toolmarks. In this regard, the analogy of fingerprints may be useful: to date, there exists no definitive proof that no two people can have identical fingerprints. Instead, the credence of fingerprint evidence rests mainly on the assertion that—across all the years in which fingerprints have been manually compared—no two people sharing the same individual prints has yet been found. The emergence of computerized image databases for fingerprints has served not only to facilitate links between pieces of evidence, but also to allow for further probing of basic assumptions. Searches across large databases of fingerprint images begin to add quantitative weight to the claim of fundamental uniqueness, and reconciliations between manual examinations and computer algorithms generate useful debates over how many specific points of similarity must be found before a match can be determined. In time, ballistic image databases may similarly be an important resource for evaluating the basic assumptions of firearms identification; however, development and study of image databases need not wait until those basic assumptions are definitively examined. Second, our work is not intended to speak to the question of whether firearms identification by a human firearms examiner can be replaced by mechanical routines. A point that we return to throughout in the report is the distinction between systems designed for search and those designed
OCR for page 20
Ballistic Imaging for verification; the two are very different, although commonly confused. NIBIN was designed as a search tool and not for verification, and as we argue, ballistic image databases are most appropriately seen as tools for search. For ballistics evidence, verification is formally made by experienced firearms examiners, who provide sworn expert testimony on evidence matches in court: hence, only direct physical examination of exhibits—and the judgment of a human firearms examiner—can certify a “hit,” or a “true” match. Our focus is on the question of whether ballistic imaging technologies perform reliably as a search tool to assist human examiners—spanning large volumes of image data and returning high-likelihood candidate matches for an examiner to consider—and not on whether computer technology can replace human examiners. Third, the proposal for this study explicitly precluded the committee from assessing the admissibility of forensic firearms evidence in court, either generally or in specific regard to testimony on ballistic imaging comparisons. We note, however, that high-subjectivity branches of forensic science are now confronting growing skepticism with regard to discernible uniqueness as a result of a number of legal and scientific studies. The standard for scientific evidence created by the U.S. Supreme Court’s decision in Daubert v. Merrell Dow Pharmaceuticals (509 U.S. 579, 1993) places high probative weight on quantifiable evidence that can be tested empirically and for which known or potential error rates may be calculated, such as identification using deoxyribonucleic acid (DNA) markers (Saks and Kohler, 2005; National Research Council, 1996). The legal context in which ballistic image evidence may be presented is too important to steer clear of entirely, and we briefly review some of the relevant legal issues and cases in Chapter 3. However, we do not in any way offer a determination of whether ballistics evidence should or should not be admissible in court proceedings. 1–A.3 Microstamping Over the course of the committee’s deliberations, debates over the feasibility of alternatives to imaging technologies that would achieve the same basic goal as a national RBID—providing an investigative lead to a point of sale—have grown in prominence. Of particular interest is the use of microstamping to directly imprint firearm parts or ammunition so that known, unique markers are imparted on bullets or cartridge casings and a connection can be made to a gun (and its point of sale) without recovery of the gun itself. The technology is also sometimes referred to as “ballistic ID tagging” or “virtual serial numbering.” Just as the issue of creating RBIDs grew in prominence when it came under serious consideration in the state of California, so, too, has legislative attention in the nation’s largest state fueled debate over microstamping. Versions of bills requiring some form
OCR for page 21
Ballistic Imaging of microstamping passed in one chamber of the California legislature (but not both) during the 2005–2006 session. However, in September 2007, AB 1471—a bill requiring microstamping on parts of new semiautomatic handguns sold in the state after January 2010—was passed by the legislature and was signed into law by the governor in October. Microstamping has grown sufficiently in stature that no bill before the 109th or (as of August 2007) 110th U.S. Congresses called for a national RBID; instead, one repeatedly introduced bill on ballistic imaging has been changed in focus to require the use of microstamping.5 Though the feasibility of a national RBID remains the committee’s primary focus, we also consider the feasibility of alternative technologies to achieve the same goal. 1–A.4 Committee Activities In carrying out its charge, the committee held six meetings beginning in February 2004, the first four of which included public sessions. Given the committee’s size and the multiple subject areas contained in its charge, the committee conducted much of its work in small working groups, including one set up to provide specific guidance to the NIST experimentation portion of the study. Committee members and staff visited local NIBIN installations at law enforcement agencies and the headquarters of Forensic Technology WAI, Inc., makers of the computer platform on which NIBIN presently operates. Committee subgroups were also permitted to perform limited experimentation using New York State’s CoBIS RBID and the ballistic image database maintained by the New York City Police Department, which is not actively linked to NIBIN but uses the same technology. To get a sense of sources of variability in bullet and cartridge markings, committee subgroups visited three firearms and ammunition manufacturers: Beretta USA, Hi-Point Firearms, and Federal Cartridge. Finally, committee subgroups examining alternative technologies visited developers of microstamping or tagging technologies for both firearms parts and ammunition. 1–B CONTEXT: THE GUN CRIME PROBLEM The primary motivation for considering the implementation of a national RBID—and for the analysis and matching of ballistics evidence, 5 See H.R. 5073, the Technological Resource to Assist Criminal Enforcement (TRACE) Act, in the 109th Congress. Introduced in April 2006, it was referred to the House Judiciary Committee but no further action was taken prior to adjournment. The same legislation was introduced in the House in the 110th Congress as H.R. 1874 on April 17, 2007, following the mass shooting at Virginia Polytechnic Institute and State University.
OCR for page 22
Ballistic Imaging in general—is to reduce gun-related crime.6 Ballistics evidence matching is intended to assist police investigations of crimes involving firearms, thereby increasing the chance of arrest, conviction, and punishment of criminals. The desired result is the incarceration of gun-using criminals and the deterrence of gun crime: if a higher percentage of such criminals are incarcerated, it may deter other criminals from using guns, and incapacitate those who get caught from committing further crimes (see, e.g., Nagin, 1998). Although gun crime constitutes a small percentage of violent crimes, guns are used in two-thirds of homicides. Criminal assaults with guns are more lethal in comparison with those involving other common weapons, and the misuse of guns by criminals creates a sense of insecurity, of “no safe place” for residents of a neighborhood where gunfire is common. The social costs of gun violence in the United States include both the direct damage of injury and death to victims and the indirect damage to the larger population whose quality of life is reduced by the threat of gun violence (Cook and Ludwig, 2000). In general, property values, business location decisions, commuting routes, and other lifestyle choices are influenced by the public’s perception of the threat of gun violence. The Washington, D.C., sniper attacks of 2002 are an extreme example that directly affected an entire metropolitan area; for some inner-city neighborhoods, gunfire is a routine occurrence that places the public in fear and distorts day to day living (Cook and Ludwig, 2002). The total annual social cost of gun crime is estimated to be $80 billion (Ludwig and Cook, 2001). Thus, tools, such as ballistic imaging technology, that can assist police in solving gun-related crime have a clear benefit for the population at large, particularly if they have some deterrent effect on gun violence. The Federal Bureau of Investigation’s Uniform Crime Reports (UCR) program provides yearly data on the number of firearm crimes known to the police. The UCR figures are based on counts of the number of murders, robberies, and aggravated assaults committed with firearms. In 2003, there were 282,641 reported total firearms crimes comprised of 137,657 (48.7 percent) robberies with firearms, 135,346 (47.9 percent) aggravated assaults with firearms, and 9,638 murders with firearms (3.4 percent).7 As shown in Figure 1-1, the yearly number of crimes committed with firearms in the United States fluctuated between 300,000 and 400,000 between 1973 and 1988, increased over the next 5 years to a peak of 581,697 in 1993, 6 The nature of gun-related crime and the adequacy of existing data related to it are comprehensively reviewed in National Research Council (2005). 7 Separate estimates for 2003 by the National Center for Injury Prevention and Control of the Centers for Disease Control and Prevention indicate approximately 47,000 nonfatal gunshot injuries for which the intent of the injury was violence-related. The data are generated from the center’s National Electronic Injury Surveillance System; see http://www.cdc.gov/ncipc/wisqars [accessed February 2008].
OCR for page 23
Ballistic Imaging FIGURE 1-1 Crimes committed with firearms, 1973–2003. SOURCE: Federal Bureau of Investigation, Crime in the United States (annually; see http://www.ojp.usdoj.gov/bjs/ [accessed February 2008]). decreased dramatically over the course of the 1990s, and remained stable through 2003. Firearms crime rates per 100,000 U.S. residents followed the same trajectory; see Figure 1-2. UCR data on the yearly number of homicides committed with firearms between 1973 and 2003 follow the same trajectory as the total number of firearms crimes per year; see Figure 1-3. However, the peaks and valleys were more pronounced in the gun homicide trend data. Gun homicides peaked in 3 years: 1974, 1980, and 1993. After 1993, there was a steep decrease to 1999. Gun homicide rates follow a similar trajectory; however, when population size is considered, the 1974 and 1993 peaks are the same (6.6 gun homicides per 100,000). The steep increase in gun homicides beginning in the 1980s and peaking in 1993 was largely driven by minority youth in urban settings (Cook and Laub, 2002; Blumstein, 1995). The youth gun violence epidemic was further concentrated among highly active criminal offenders who tended to be involved in street gang or illegal drug activity (Braga, 2003; Kennedy et al., 1996). In most cities, gun violence problems remain concentrated among a small number of criminally active youth who are involved in gangs or criminal groups (Braga et al., 2002). Cities vary widely in the amount of gun crime they experience and
OCR for page 24
Ballistic Imaging FIGURE 1-2 Firearms crime rates, 1973–2003. SOURCE: Federal Bureau of Investigation, Crime in the United States (annually; see http://www.ojp.usdoj.gov/bjs/ [accessed February 2008]). FIGURE 1-3 Homicides committed with firearms, 1973–2003. SOURCE: Federal Bureau of Investigation, Crime in the United States (annually; see http://www.ojp.usdoj.gov/bjs/ [accessed February 2008]).
OCR for page 25
Ballistic Imaging the numbers of crime guns local police departments recover. Table 1-1 presents gun homicide counts and rates, as well as the number of crime guns recovered, for 32 cities that participated in ATF’s Youth Crime Gun Interdiction Initiative (YCGII) in 2000 and were judged by ATF to be submitting all recovered firearms for tracing (U.S. Bureau of Alcohol, Tobacco, TABLE 1-1 Gun Homicides and Crime Gun Recoveries in 32 Cities in 2000 City Gun Homicides Homicide Rate per 100,000 Gun Recoveries Atlanta, GA 108 25.9 1,141 Baltimore, MD 202 31.0 4,295 Baton Rouge, LA 33 14.5 1,068 Boston, MA 26 4.4 896 Camden, NJ 17 21.3 165 Charlotte, NC 57 9.1 2,041 Chicago, IL 415 14.3 8,570 Cincinnati, OH 7 2.1 877 Dallas, TX 177 14.9 3,005 Gary, IN 56 54.5 792 Houston, TX 165 8.4 3,909 Indianapolis, IN 67 8.4 3,592 Los Angeles, CA 430 11.6 3,877 Louisville, KY 33 12.9 1,637 Memphis, TN 110 16.9 3,244 Milwaukee, WI 90 15.1 2,283 Minneapolis, MN 38 9.9 949 Nashville, TN 56 10.5 2,297 New Orleans, LA 175 36.1 1,965 New York, NY 434 5.4 6,284 Newark, NJ 40 14.6 584 Oklahoma City, OK 24 4.7 856 Philadelphia, PA 259 17.1 3,041 Phoenix, AZ 110 8.3 4,778 Piedmont Triad, NCa 38 7.7 699 Portland, OR 14 2.6 857 Richmond, VA 53 26.8 1,109 Salinas, CA 16 10.6 327 San Antonio, TX 45 3.9 1,294 San Jose, CA 8 0.9 1,476 St. Louis, MO 90 25.8 2,612 Tucson, AZ 49 10.1 2,135 aGreensboro, High Point, and Winston-Salem, NC. SOURCES: Gun homicide data from FBI Supplementary Homicide Reports, 2000; see http://www.icpsr.umich.edu [accessed February 2008]. Gun recovery data from U.S. Bureau of Alcohol, Tobacco, and Firearms, 2002.
OCR for page 26
Ballistic Imaging and Firearms, 2002; see Box 9-1). Not surprisingly, large cities—New York, Los Angeles, and Chicago—report the largest numbers of gun homicides. However, there are smaller cities that experience dramatically higher rates of gun violence relative to the large cities. Gary, Indiana, had the highest rate of gun homicides per 100,000 residents with 54.5, followed by New Orleans with 36.1 percent, and Baltimore with 31.0 percent. Because of the variability in gun crime rates by locality, different localities may have different baseline needs for ballistic imaging technology (and, potentially, different levels of benefit from its refinement). 1–C BALLISTIC IMAGING, FIREARMS IDENTIFICATION, AND “BALLISTIC FINGERPRINTING” Analysis of ballistics evidence may provide a link between two shooting incidents if it is determined that the same weapon was fired in both. That information may be helpful to investigators since it suggests that the incidents involved the same shooter, or involved two shooters who were linked by the transfer of the gun in question. Alternatively, the ballistics evidence match can provide a link between a shooting incident and a particular gun, perhaps one that has separately been found and placed in police custody; this information may be helpful to the investigation if the identity of the owner or possessor of that gun is known or could be determined through further investigation. It is important to clarify several terms and the distinctions among them. First, ballistic imaging is not identical to firearms identification. Traditional firearms identification techniques, relying on the direct viewing of specimens under a comparison microscope by a trained firearms examiner, have been used in investigations for decades. As discussed in Section 1–A.2, the identification and confirmation of fired bullets or cartridge cases as having been fired from a specific firearm is the responsibility of human examiners. Ballistic imaging is a means of searching across a large number of exhibits—in greater numbers and across broader expanses of geography than a human examiner could possibly achieve—to suggest possible matching candidates. Ballistic imaging would more accurately be described as a form of computer-assisted firearms identification. The unique innovation that ballistic imaging technology has added to the field is the “cold hit”—the generation of possible links between specimens and cases arising only through querying a database. A cold hit can be particularly valuable for furthering the investigation of shooting crimes that lack an obvious suspect or even any clear leads. Research on police clearance of homicide cases (Wellford and Cronin, 1999, 2000) suggests that the availability of witnesses (who can identify the offender or victim and who may be able to suggest the whereabouts of the offender) and swift
OCR for page 27
Ballistic Imaging action by the first officers on the scene are major contributors to success in closing a case. By comparison, crimes for which eyewitness testimony is not available (or where witnesses may be unwilling to come forward), such as is common in drug-related homicides, are harder to solve. In cases for which witnesses and early data on possible suspects are lacking, a cold hit on ballistics evidence generated through a routine database search could provide an important investigative lead. Generally, ballistic imaging offers the opportunity for more rapid searching across a high volume of candidates than is possible using conventional techniques. In traditional firearms identification, a firearms examiner’s cognitive task in examining specimens under a comparison microscope is to form a mental pattern of identifying marks and features on bullet and cartridge case evidence and to match that pattern against those from other exhibits. Accordingly, searching through large amounts of ballistic evidence and verifying a match can be a very labor-intensive and time-consuming task. Making connections between different cases relies on the visual memory of the firearms examiner or—if all exhibits are not viewed and remembered by the same person—recognition of features from photographs in open case files or posts on bulletin boards. Ballistic imaging technology allows images of bullets or casings to be cataloged, indexed, scored, and ranked. A firearms examiner can visually compare high-ranked pairs of images on the screen, much as a radiologist might read a digital mammogram or other X-ray, and the physical evidence items for promising matches can then be requested as appropriate for confirmation. The general ballistic imaging methodology we describe in this study has been popularly referred to as ballistic fingerprinting, a term that carries both positive and negative connotations and that is misleading in a very important sense. Most commonly used in relation to a national reference ballistic image database, with the idea of logging a newly sold gun’s “fingerprint” before or as a condition of sale, “ballistic fingerprinting” naturally suggests a connection to the more widely known practice of recording human fingerprints. What is fundamentally misleading about equating “ballistic imaging” and “ballistic fingerprinting” is the point of reference—a human fingerprint is an attribute of that human, and a determined match between a latent fingerprint found at a crime scene and a fingerprint in police files suggests a direct connection between a crime and a suspect. However, the markings imparted to fired bullets and casings are attributes of a firearm, not the person who fires it.8 8 Though “ballistic fingerprinting” has become a popular term in recent years, references to ballistic toolmarks as mechanical fingerprints date back to the formative days of firearms identification. Hatcher (1935:265, 275), one of the seminal texts in the field, notes that “these [toolmarks] are what might very aptly be described as the ‘finger prints of the firing pin and
OCR for page 28
Ballistic Imaging Absent other evidence, firearms identification and ballistic imaging do not automatically generate a mapping from ballistics evidence to a possible perpetrator. We return to this point in Section 3–A, but it is important to note here that fingerprint (and DNA) evidence refer to attributes of a particular person, but they do not necessarily point to that person as the criminal offender. That is, the presence of this evidence can place a person at the location of a crime, but not necessarily demonstrate that they were there at the time of the crime or that they committed the act in question. The intent of a national RBID is to provide a relatively quick connection between recovered ballistics evidence and a point of sale. However, additional work from a national RBID “hit” would still be necessary to derive a person’s name from the point of sale and that this person—the original purchaser of the firearm—is not necessarily the person who used the gun in crime. 1–D OVERVIEW OF REPORT Part I of this report describes the context for ballistic image analysis. Chapter 2 describes the toolmarks imparted on bullets and cartridge casings as a result of firing, reviewing the sources of variability inherent in the manufacture of firearms and in the process of firing a gun. Chapter 3 describes the nature of ballistics evidence in more detail, focusing on traditional firearms identification techniques and the studies that have been performed on the uniqueness and reproducibility of firearms-related toolmarks as discerned using conventional microscopy. Part II deals with the current state of ballistic imaging and the existing national image database, NIBIN. Chapter 4 discusses the technology used for acquiring images and scoring and ranking them, focusing on the IBIS platform used by the NIBIN program. Chapter 5 describes the evolution of the NIBIN program and its structure and summarizes what is known about the NIBIN system’s performance. Drawing from both these chapters, Chapter 6 outlines operational and technical enhancements that could improve NIBIN. Part III addresses the basic titular charge to the committee, describing evidence on variability in ballistics evidence and the implications for a national reference ballistic image database. Chapter 7 introduces a major technical enhancement that the committee chose to explore as an option breech block on the primer.’” If all the gross, class marks are the same between two bullets, “this does not, however, prove in any way that [a suspect bullet] came from that particular gun as there are hundreds or even thousands of guns of each type manufactured.… Fortunately, however, each and every barrel has its own ‘finger prints’ which it leaves on a bullet, and identification by these marks is just as certain as identification of a criminal by his finger prints.”
OCR for page 29
Ballistic Imaging for either the current NIBIN program or a wide-scale national reference database. That enhancement is the replacement of two-dimensional photography with three-dimensional topography, and we briefly describe that technology along with historical alternatives to photography in firearms analysis. Chapter 8 reviews the experimental efforts conducted by NIST in support of the committee’s work, as well as limited experimental work using the New York State CoBIS database. Chapter 9 builds from the new experimental evidence and from studies (described in Chapter 4 and elsewhere) in articulating the arguments associated with creating a national reference database. Part IV, on future directions, begins in Chapter 10 by discussing alternative technologies to achieve the same goal as a national reference ballistic image database. In particular, we review proposals to microstamp firearms parts or individual pieces of ammunition with unique etched identification codes. Chapter 11 closes the report with general guidance on the process of developing systems for image search, retrieval, processing, and scoring, suggesting “best practices” for development of any such program (whether advancing current two-dimensional photography techniques or changing to three-dimensional topography). Appendix A offers additional detail on the use of ballistic imaging technology in Boston, one locale where the current NIBIN system appears to be well used and well supported.