You've probably seen it on TV: To determine whether a single weapon was used in multiple crimes, forensics labs run images of a spent bullet through a database and look for "hits," or matches with other bullets. That database has a real-life analog, and it is maintained by the National Integrated Ballistics Information Network (NIBIN), a collection of about 150 specialized systems in U.S. federal, state, and local crime labs.
The NIBIN database has yielded more than 68,000 hits since the image collection project began in the 1990s, according to the Bureau of Alcohol, Tobacco, Firearms, and Explosives (ATF) website – 68,000 potential matches between bullets or cartridge cases, which specialists have then confirmed with manual inspections.
"The NIBIN database is a powerful way to link firearms, suspects, and crimes," says Theodore Vorburger of the Engineering Physics Division in NIST's Physical Measurement Laboratory (PML).
NIBIN labs all use a common system for imaging and correlation called IBIS (Integrated Ballistics Identification System). But in order for the database to make these comparisons, it's vital that the microscope images be acquired consistently, Vorburger continues. Photos collected on different days by different labs might have achieved different levels of contrast, or different intensity in lighting, "whereas you want to be pretty sure that if you measure a cartridge today and then measure it tomorrow, or six months from now, you will get the same image," Vorburger says.
That's where NIST comes in. To ensure that the IBIS technicians' imaging techniques were consistently gathering good imaging data, NIST PML staff were tasked by the ATF in the mid-2000s to design a system that NIBIN labs could use to calibrate their imaging machines. Accordingly, NIST developed a "standard bullet" and a "standard cartridge case," with markings that were representative of what you would see on a real spent bullet or cartridge case, but with a surface topography that was known to a high degree. Master images of these standard objects were taken at the ATF National Laboratory Center.
The idea was that a NIBIN lab could, at regular intervals, take an image of a standard bullet or cartridge case with their IBIS system and compare that image to the master image acquired at ATF. Comparing the two would give them a correlation score; a sufficiently high score would assure them that they were maintaining high quality control.
Then, in the late 2000s, NIST was tasked with determining how high a correlation score had to be for acceptability. PML staff conducted a study with the participation of 19 experts from 13 crime labs and created a chart that NIBIN labs could use with the IBIS imaging model current at the time.
"But then IBIS came out with a new model," Vorburger says.
Now, PML researchers have released a new downloadable tool, an interactive "control chart" that will help forensics specialists control the quality of images using this newest model of imager, called the BRASSTRAX™ machine. Essentially a black box containing a camera and lights, the new imager is capable of double the pixel resolution of the previous system, has 3D capabilities,** and is strictly for microscopic measurement of spent cartridge cases.
Just as they had for the previous study, the PML team was tasked with finding out how high the correlation scores needed to be to ensure that images were "good enough" for the database.
The BRASSTRAX™ system takes microscope images of three key markings typically found on spent cartridge cases. One is an impression made by the firing pin, which strikes the cartridge's primer and ignites the gunpowder, creating an explosion that propels the bullet out of the barrel.
Another is the so-called breech face marking, a roughness profile imprinted on the head of the cartridge when recoil thrusts it against the back inside wall of the gun; the breech face impression becomes "a kind of fingerprint" for the weapon, explains Vorburger's PML collaborator Jun-Feng Song.
Finally there is the ejector mark, made at the conclusion of the firing process, when the spent cartridge case is extracted and ejected from the gun to make room for a new cartridge.
In their assessment, PML researchers asked 14 ballistics examiners using BRASSTRAX™ machines at 11 different NIBIN laboratories to take images of NIST's Standard Cartridge Case (SRM 2461), at regular intervals, over the course of about a year. These images were compared to the master images of the standard cartridges taken by ATF with their own BRASSTRAX™ system. A correlation score was assigned to each of the NIBIN images depending on how well it matched the masters.
The NIST team found higher correlation scores on average with photos taken using the new system than with the previous imager. For the markings made by the breech face and the ejector, these increases were fairly small. But for the comparatively tiny imprints made by the firing pin, correlations were significantly higher, allowing PML researchers to recommend a much higher limit for acceptable images of firing pin impressions.
Forensics specialists using the new interactive control chart produced by NIST's PML can enter data about their images into a spreadsheet and get immediate feedback about whether their images are within acceptable quality limits. The PML team was also able to set uncertainty levels for the firing pin and breech face markings, which investigators can use to determine their confidence levels about a match.
Ultimately, the researchers hope that this new control chart will help IBIS technicians find more database hits, Vorburger says.
The work was supported by the National Institute of Justice and by the Law Enforcement Standards Office under the NIST Special Projects Office.
— Reported and written by Jennifer Lauren Lee
* Details of the investigation, including a link to the control chart, are available here. T.V. Vorbuger, J.H. Yen, J-F Song, R.M. Thompson, T.B. Renegar, Z.A. Zheng, M. Tong, M.G. Ols (2015). The Second National Ballistics Imaging Comparison (NBIC-2). Journal of Research (NIST JRES), 219, 644-673. DOI: http://dx.doi.org/10.6028/jres.119.028.
** Since BRASSTRAX™ images use the same correlation software as the older imaging system, photos taken with both systems can be compared in the database.
Any mention of commercial products within NIST web pages is for information only; it does not imply recommendation or endorsement by NIST.