10
Automated Fingerprint Identification Systems

In the late 1970s and early 1980s law enforcement agencies across the Nation began adopting Automated Fingerprint Identification Systems (AFIS) to improve their efficiency and reduce the amount of time it took to identify (or not exclude) a given individual from a fingerprint or to conduct a background investigation. AFIS introduced an enormous improvement in the way local, state, and federal law enforcement agencies managed fingerprints and identified people. Before the use of AFIS, the fingerprint identification process involved numerous clerks and fingerprint examiners sifting through thousands of tediously classified and cataloged paper fingerprint cards, while dealing with delays and challenges caused by the realities of exchanging information with other agencies by mail, fax, or other means. With AFIS, fingerprint examiners use computer workstations to mark the features of a scanned fingerprint image (e.g., ridge endings, bifurcations), encode the resulting data in a machine-readable format, and then search for similar fingerprints in an associated database of known fingerprints and records. AFIS searches are fast, and they often allow examiners to search across a larger pool of candidates. Although challenging cases can be time consuming, depending on the size of the database being searched and the system’s workload, AFIS often can return results to the examiner within minutes.

AFIS searches today fall into two distinct categories:

10-print searches, which typically involve comparing relatively high-quality, professionally obtained fingerprint images—for example, prints taken during an arrest or booking or as part of a background check—



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 269
10 Automated Fingerprint Identification Systems In the late 1970s and early 1980s law enforcement agencies across the Nation began adopting Automated Fingerprint Identification Systems (AFIS) to improve their efficiency and reduce the amount of time it took to identify (or not exclude) a given individual from a fingerprint or to conduct a background investigation. AFIS introduced an enormous improvement in the way local, state, and federal law enforcement agencies managed finger- prints and identified people. Before the use of AFIS, the fingerprint identi- fication process involved numerous clerks and fingerprint examiners sifting through thousands of tediously classified and cataloged paper fingerprint cards, while dealing with delays and challenges caused by the realities of exchanging information with other agencies by mail, fax, or other means. With AFIS, fingerprint examiners use computer workstations to mark the features of a scanned fingerprint image (e.g., ridge endings, bifurcations), encode the resulting data in a machine-readable format, and then search for similar fingerprints in an associated database of known fingerprints and records. AFIS searches are fast, and they often allow examiners to search across a larger pool of candidates. Although challenging cases can be time consuming, depending on the size of the database being searched and the system’s workload, AFIS often can return results to the examiner within minutes. AFIS searches today fall into two distinct categories: 10-print searches, which typically involve comparing relatively high- quality, professionally obtained fingerprint images—for example, prints taken during an arrest or booking or as part of a background check— 

OCR for page 269
0 STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES with fingerprint records in an agency database, such as the FBI’s Inte- grated Automated Fingerprint Identification System (IAFIS) or a state’s criminal fingerprint database; and Latent print searches, which are considerably more complicated than 10-print searches. In a latent print search, a fingerprint examiner at- tempts to identify an individual by comparing a full or partial latent fingerprint from a crime scene with the records contained in an AFIS database. Latent prints are regularly of poor quality and may be only a partial print, and often fingerprint examiners may not even know from which finger a given latent print came. A third category (albeit one that includes elements of both categories listed above) might also be called “unidentified burned, decomposed, or fragmented prints,” which may be either a complete 10-print card to be compared with known prints on file to confirm identity or partial prints recovered from the skin or dermis of damaged fingers of an unknown de- cedent to determine identity. This third category can include prints from single individuals recovered from a small single event or victims of a mass casualty event resulting from naturally occurring catastrophes or terrorism. In either case, AFIS systems have reduced the time required to accomplish many identifications from weeks to hours. Today, the process of populating AFIS systems with records is man- aged primarily by uploading 10-print records from police bookings and background checks. Because images from these sources are generally of good quality (indeed, poor-quality 10-print records are normally redone at the time they are taken), an automated algorithm is adequate for extract- ing the features used to index an image for retrieval. Computer algorithms work well for performing comparisons of 10-print records (e.g., to see if the prints taken when one applies for a security clearance match the prints taken during a previous background check). However, submitting a latent print for comparison is a more customized process, requiring fingerprint examiners to mark or adjust the features manually to retrieve stored prints with the same features in analogous places. Because latent print images normally are not as clear or as complete as images from a 10-print card, the image processing algorithms used for 10-prints are not as good as the human eye in spotting features in poor images. AFIS has been a significant improvement for the law enforcement com- munity over the past decades, but AFIS deployments today are still far from optimal. Many law enforcement AFIS implementations are stand-alone systems or are part of relatively limited regional networks with shared databases or information-sharing agreements—the Western Identification

OCR for page 269
 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS Box 10-1 The Western Identification Network WIN was formed in May 1988 to facilitate the creation of a multistate AFIS implementation. A year later, the state legislatures of Alaska, California, Idaho, Oregon, Nevada, Utah, Washington, and Wyoming appropriated the necessary funding to begin work on the system. The initial WIN AFIS was installed in Sacramento, California, with remote subsystems in Cheyenne, Wyoming; Salt Lake City, Utah; Boise, Idaho; Carson City, Nevada; and Salem and Portland, Oregon. Booking terminals also were installed in numerous locations throughout these states, and existing similar stand-alone systems in Alaska, California, and Washington were connected to WIN in 1990 to complete the initial network. At first, WIN’s centralized automated database included 900,000 fingerprint records, but after connecting to Alaska, California, and Washington, the number of searchable fingerprint records in- creased to more than 14 million. Today, WIN members have access to more than 22 million fingerprint records from the western United States. NOTE: For information about WIN, see www.winid.org/winid/who/documents/WINService StrategyJanuary2008.pdf. Network (WIN) is one example of such a regional network (for more in- formation on WIN, see Box 10-1). Today, AFIS systems from different vendors most often cannot interop- erate with one another. Indeed, different versions of similar systems from the same vendor sometimes cannot share fingerprint data with one another. In addition, many law enforcement agencies also access the FBI’s IAFIS da- tabase1 through an entirely separate stand-alone system—a fact that often forces fingerprint examiners into entering fingerprint data for one search multiple times (at least once for each system being searched). There is no doubt that much good work has been done in recent years aimed at improving the interoperability of AFIS implementations and da- tabases (see Box 10-2), but the committee believes that, given the potential benefits of more interoperable systems, the pace of these efforts to date has been too slow, and greater progress needs to be made toward achieving meaningful, nationwide AFIS interoperability. 1 See www.fbi.gov/hq/cjisd/iafis.htm.

OCR for page 269
 STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES Box 10-2 Working Toward AFIS Interoperability As early as 1986, the American National Standards Institute (ANSI) and the National Bureau of Standards (now known as the National Institute of Standards and Technology, or NIST) were working on ways to facilitate the exchange of fin- gerprint data. Their collaboration produced a standard defining minutiae data and both low- and high-resolution fingerprint images. The standard was not successful, however, because of conflicts with proprietary systems. In 1993, ANSI and NIST teamed up again to create another fingerprint data standard, a standard later updated in 1997. It defined standards for minutiae data and low- and high-resolution fingerprint images in both binary and grayscale for- mat, as well as methods for compressing and decompressing image data. In the late 1990s, the International Association for Identification’s AFIS Com- mittee successfully demonstrated a method of conducting remote fingerprint searches across jurisdictions and across equipment from different vendors.a In 2003, the ANSI/NIST standard was updated again. It grew to include 16 record types in total, with the addition of standards for such things as palm print data and latent print data.b The standard was recently updated once more and has subsequently been approved by ANSI’s Board of Standards Review as an ANSI standard.c The NIST-sponsored Minutiae Interoperability Exchange Test (MINEX) pro- gram is an ongoing series of coordinated development efforts aimed at improving the performance and interoperability of fingerprint minutiae standards. In 2004, the original project undertook to determine the feasibility of using minutiae data (rather than image data) as the interchange medium for fingerprint information between different fingerprint matching systems.d a The committee’s final report is available at www.onin.com/iaiafis/IAI_AFIS_071998_Report. pdf. b For more information on the ANSI/NIST standards, see P. Komarinski. 2005. Automated Fingerprint Identification Systems. Boston: Elsevier Academic Press, pp. 162-166. c This approved revision of the ANSI/NIST-ITL 1-2000 standard is now available as NIST Special Publication 500-271: Data Format for the Interchange of Fingerprint, Facial, & Other Biometric Information-Part 1 (ANSI/NIST-ITL 1-2007) at http://fingerprint.nist.gov/standard/ Approved-Std-20070427.pdf. d More information about the work of the MINEX series is available at http://fingerprint.nist. gov/minexII/. INTEROPERABILITy CHALLENgES Despite the work done to date to achieve broader AFIS interoperability and its potential benefits (i.e., more crimes solved, quicker and more effi-

OCR for page 269
 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS cient searches, and better use of limited law enforcement resources), several persistent challenges to reaching this goal remain. Technical Challenges The technical challenges to AFIS interoperability involve both those that are encountered and addressed by the information technology commu- nity in other disciplines (such as data sharing and algorithmic performance) and those that are specific to AFIS and the sharing of fingerprint informa- tion (e.g., feature identification, reliability of latent print comparisons). In addition, systems will need to be designed with the flexibility to handle other kinds of biometric data in the future (e.g., iris and palm scans and possibly genomic data). As these latter challenges are addressed, retrieval algorithms within proprietary AFIS systems also may tend to converge, which could simplify the broader interoperability challenges. Creating useful technical standards is never a simple undertaking, es- pecially given a diverse array of stakeholders, proprietary systems, and ever-advancing technological capabilities (e.g., improved pattern recogni- tion, better hardware, increased data compression). However, the successful interoperability of other distributed information networks—such as modern banking systems (e.g., ATM machines2), information sharing networks in the real estate world,3 the Centers for Disease Control and Prevention’s Public Health Information Network,4 and even the Internet itself, each of which functions only by reliance on a number of finely crafted and agreed standards and protocols—is proof that efforts to develop and implement standards pay off in the end by allowing greater collaboration and sharing of information. One other major area of technical challenge to achieving AFIS interop- erability involves the algorithms that systems use to identify features in fin- gerprint images (e.g., how a system determines that a given pattern of pixels corresponds to a true ridge ending or bifurcation and how it infers what type of feature those pixels actually represent). To date, these algorithms 2 Indeed, financial card transactions are facilitated by their own ISO standard (ISO 8583-1:2003). For more information, see www.iso.org/iso/iso_catalogue/catalogue_tc/ catalogue_detail.htm?csnumber=31628. 3 See, e.g., the Metropolitan Regional Information System (MRIS) at www.mris.com/about/ WhoWeAre.cfm. 4 CDC’s Public Health Information Network is a national initiative to improve the capacity of the public health community to use and exchange information electronically by promoting the use of standards and defining functional and technical requirements. The network employs a messaging system (PHINMS) to rapidly and securely share sensitive health information among CDC and other local, state, and federal organizations over the Internet—information such as HIV records, pandemic information, and information on bioterrorism. Complete information about PHIN and PHINMS is available at www.cdc.gov/phin/.

OCR for page 269
 STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES have been largely proprietary and vendor specific (i.e., different for each type of system). In fact, experienced latent print examiners have found that different systems will retrieve different stored prints in response to a given input map of features, and they have learned system-specific ways of an- notating features on a latent print in order to maximize the success of each system’s (inferred) search algorithms. However, achieving broad-based AFIS interoperability will require baseline standards for these algorithms, so that fingerprint examiners can be assured of consistent feature mapping across systems. As mentioned previously, fingerprint examiners have learned by experience to provide different inputs to different vendors’ systems, often purposely leaving out information—knowing that the added input will degrade the search quality: The examiner does not necessarily encode every point he can find in the latent print. LPU [latent print unit] examiners have learned through ex- perience with the IAFIS program which types of points are most likely to yield a correct match. LPU Unit Chief Meagher told the OIG [Office of Inspector General] that examiners are taught to avoid encoding points in areas of high curvature ridge flow, such as the extreme core of a print. Unit Chief Wieners and Supervisor Green told the OIG that IAFIS does not do well when asked to search prints in which points have been encoded in two or more clusters separated by a gap. One reason is that IAFIS gives significant weight to the ridge count between points. If the ridge count between two clusters of points in a latent is unclear, IAFIS may fail to retrieve the true source of the print. Thus, an examiner will not necessar- ily encode every point that can be seen in a latent fingerprint, but rather may limit his encoding to points in a defined area in which the ridge count between points is clear.5 The fact that today’s systems often do not effectively utilize most of the available feature information and require substantial input from fingerprint examiners suggests that there is significant room for improvement. An ideal, comprehensive AFIS, for example, would be capable of automated: • reading of latent prints; • encoding of most features of usable quality, including those fea- tures identified as Level 1 (fingerprint classes such as whorl, arch), Level 2 (minutiae), Level 3 (pores, cuts), and ridge paths, together with a provision for including other features that could be defined by the vendor/user; 5 Office of the Inspector General, Oversight and Review Division, U.S. Department of Jus- tice. 2006. A Reiew of the FBI’s Handling of the Brandon Mayfield Case, p. 119.

OCR for page 269
 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS • recognizing absent, blurred, double/multioverlap, poor-quality sec- tions of an observed print and encoding the system to downweight, or omit entirely, during the search process; • recognizing any orientation information; • conducting database searches; • providing “best matches”; and • collecting statistical data based on the quality of the print and numbers/types of features. Other technical challenges might include the development and use of a secure Web interface (or an analogous system) that would permit autho- rized latent print examiners in any jurisdiction to submit queries to IAFIS and other federated AFIS databases, as well as the development of standard procedures for maintaining AFIS databases securely, removing redundan- cies, ensuring that fingerprint data are entered properly, and conducting quality control and validation of searches (i.e., ensuring that queries are actually searching an entire database). Although some of the capabilities mentioned here are present in today’s commercially available systems, sig- nificant improvement still can be realized. Support from Policymakers Given the complexity of the AFIS interoperability challenge and the large number of players whose contributions and cooperation will be nec- essary to meet that challenge, it is clear that no effort aimed at nation- wide interoperability will succeed without strong, high-level support from policymakers in federal and state government. Resources available to law enforcement agencies for the deployment, use, and maintenance of AFIS systems vary greatly from jurisdiction to jurisdiction, and the considerable expenses associated with purchasing, maintaining, training for, operating, and upgrading an AFIS implementation—which can easily cost millions of dollars6—must be well thought out and weighed against other competing costs and interests facing law enforcement. The committee hopes that this report will help convince policymakers of the benefits to nationwide interoperability and move them to provide much-needed support to law enforcement agencies, vendors, and research- ers to help them achieve this goal. Indeed, the committee believes that true AFIS interoperability can be achieved in a timely manner only if policymak- ers provide a strong, clear mandate and additional funding from federal and state governments—both to support the research and development 6 See P. Komarinski. 2005. Automated Fingerprint Identification Systems. Boston: Elsevier Academic Press, p. 145.

OCR for page 269
 STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES work necessary to achieve truly interoperable systems and to assist law enforcement agencies in purchasing, implementing, and managing systems and training personnel. Vendors As suggested above, AFIS equipment and service vendors must coop- erate to ensure nationwide AFIS interoperability. However, to date—and as one could reasonably expect in a technology sector in which product differentiation and the maintenance of competitive advantages are prime concerns—vendors have had little incentive to design their systems to en- able them to share information with competitors’ systems. The committee believes that increased cooperation among AFIS vendors is a key to achiev- ing meaningful interoperability. For example, one can imagine how it might prove useful if AFIS vendors could collaborate (perhaps through work facilitated by the proposed National Institute of Forensic Science [NIFS]) on developing standard (or baseline) retrieval algorithms. Such a step con- ceivably could make it less time consuming for fingerprint examiners to run searches on many different systems because they would not have to manu- ally tune their searches to work on the systems of different vendors. Administrative, Legal, and Policy Issues As noted earlier, most AFIS implementations are either stand-alone systems or are part of relatively limited regional databases. To achieve truly interoperable systems, jurisdictions must work more closely together to craft acceptable agreements and policies to govern the routine sharing of fingerprint information. NIFS can facilitate the development of standard agreements along these lines, which could include issues such as the extent of system access to other jurisdictions, the management of search priorities, and the recovery of costs associated with processing the requests from out- side agencies. In addition, many jurisdictions also might want assurances that they will not be held responsible for any possible misuse of fingerprint information that is provided to other law enforcement agencies. CONCLuSIONS AND RECOMMENDATION Great improvement is possible with respect to AFIS interoperability. Many crimes no doubt go unsolved today simply because investigating agencies cannot search across all the individual databases that might hold a suspect’s fingerprints or contain a match for an unidentified latent print from a crime scene. It is possible that some perpetrators have gone free because of the limitations on fingerprint searches.

OCR for page 269
 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS The committee believes that, in addition to the technical challenges noted above, a number of other critical obstacles to achieving nationwide AFIS interoperability exist involving issues of practical implementation. These include (1) convincing federal and state policymakers to mandate nationwide AFIS interoperability; (2) persuading AFIS equipment vendors to cooperate and collaborate with the law enforcement community and re- searchers to create and use baseline standards for sharing fingerprint image and minutiae data and interfaces that support all searches; (3) providing law enforcement agencies with the resources necessary to develop interoper- able AFIS implementations; and (4) coordinating jurisdictional agreements and public policies that would allow law enforcement agencies to share fingerprint data more broadly. Given the disparity in resources and information technology expertise available to local, state, and federal law enforcement agencies, the relatively slow pace of interoperability efforts to date, and the potential gains that would accrue from increased AFIS interoperability, the committee believes that a new emphasis on achieving nationwide fingerprint data interoper- ability is needed. Recommendation 12: Congress should authorize and appropriate funds for the National Institute of Forensic Science (NIFS) to launch a new broad-based effort to achieve nationwide fingerprint data interoperability. To that end, NIFS should convene a task force comprising relevant experts from the National Institute of Standards and Technology and the major law enforcement agencies (including representatives from the local, state, federal, and, perhaps, international levels) and industry, as appropriate, to develop: (a) standards for representing and communicating image and minutiae data among Automated Fingerprint Identifica- tion Systems. Common data standards would facilitate the sharing of fingerprint data among law enforcement agencies at the local, state, federal, and even international levels, which could result in more solved crimes, fewer wrongful identifications, and greater efficiency with respect to fingerprint searches; and (b) baseline standards—to be used with computer algorithms— to map, record, and recognize features in fingerprint images, and a research agenda for the continued improve- ment, refinement, and characterization of the accuracy of these algorithms (including quantification of error rates).

OCR for page 269
 STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES These steps toward AFIS interoperability must be accompanied by the provision of federal, state, and local funds to support jurisdictions in up- grading, operating, and ensuring the integrity and security of their systems; the retraining of current staff; and the training of new fingerprint examiners to gain the desired benefits of true interoperability. Additionally, greater scientific benefits can be realized through the availability of fingerprint data or databases for research purposes (using, of course, all the modern security and privacy protections available to scientists when working with such data). Once created, NIFS might also be tasked with the maintenance and periodic review of the new standards and procedures.