Is Biometrics Feasible in Correctional Facilities?

Keeping track of inmates within a prison or jail is a constant challenge, especially as they move from one part of the facility to another. Monitoring their movements requires corrections officers to accurately identify individual prisoners by sight as they pass through security posts. It also requires frequent telephone and radio communications between officers at two or more security posts, paper passes authorizing inmates’ movements, and dry-erase or clipboards with handwritten records to note when prisoners leave one area and enter another. Despite the best precautions and well-thought-out practices, mistakes can be made, officers’ attention can be diverted, and late-arriving inmates not noticed or searched for promptly.


Late-arriving, out-of-place prisoners can cause problems in correctional settings. If nothing else, it means that prison staff do not know where a particular inmate is at any given time. That prisoner may simply have stopped to chat with friends. Or, more seriously, he or she may be engaging in illegal activities. Assaults and even murders have been committed by inmates as they moved from one part of a prison or jail to another.


In an effort to improve how inmate movements are tracked within prisons and jails, the National Institute of Justice has been testing the use of biometrics at the U.S. Naval Consolidated Brig in Charleston, S.C. The $1 million technology demonstration project is a joint effort of NIJ, the U.S. Navy’s Space and Naval Warfare Systems Center, the Charleston brig, and the U.S. Department of Defense’s Biometrics Management Office.


Biometrics has been used previously to track the movement of staff, visitors, and prisoners in and out of correctional facilities. It has also been used to account for staff members in the event of a riot or other prison disturbance. This project represents the first use of biometrics to track prisoner movements within a prison or jail. It was designed to employ computer-based methods of tracking inmates to improve the efficiency of corrections specialists and brig officials and to demonstrate how advanced technology can make corrections facilities safer.


The term “biometrics” refers to a variety of methods to verify a person’s identity using physiological or behavioral characteristics such as iris, retinal, and facial recognition; hand and finger geometry; fingerprint and voice identification; and dynamic signature. It has the advantage of not requiring a person to remember a user name, password or series of numbers while confirming that the person is who he or she claims to be. Practical uses of biometrics include allowing persons access to keyless cars, rooms, and buildings; to financial and other personal accounts; and to the departure areas of airport terminals. More broadly, it is used to prevent identity theft, preserve the confidentiality of information and reduce fraud.


Biometric systems can use several different physical and/or behavioral characteristics for identification and verification. Some are more technologically and commercially advanced than others. Determining which biometric method to employ depends on how the system is to be used, the level of accuracy and reliability required, and other factors such as cost and speed. Biometric methods can also vary significantly from one application to another and even from one vendor to another.


Biometrics systems are usually deployed using a three-step process. First, a camera, scanner or other sensor takes an image or picture. Second, that image is made into a pattern called a biometric signature. For example, with fingerprints the signature comprises minutia points along a finger’s ridges, splits and end lines. Voice recognition involves patterns of cadence, pitch and tone. Hand and finger geometry measures physical characteristics such as length and thickness.


Third, the biometric signature is converted into a template using a mathematical algorithm. Templates contain biometric and other data in the form of numbers that are either embedded on a plastic card or stored in a database. Some systems use a card that can be inserted in or held near a scanner that feeds the information on the card into a computer. Other systems do not require a card; they simply scan the biometric data. In either system, the computer compares the biometric signature captured by the scanner with those already in its files to find the correct or closest match.


NIJ and DoD began examining biometric techniques for criminal justice purposes in 2000. As part of that effort, NIJ and DoD identified the Charleston naval brig as a demonstration site. The brig is a relatively small, well-managed jail with approximately 400 mostly low-risk prisoners. The Navy wanted to upgrade security at the brig and make it more efficient. At the same time, the adjacent Space and Naval Warfare Systems Center was available to help develop the biometrics system and the computer software necessary to run it.


Identifying Inmates


Called the Biometric Inmate Tracking System, the project was implemented in phases that, together, transformed the existing manual system into a computer-based system and then into a biometric- and computer-based system. In carrying out that transformation, project designers had to find the biometric method that would work best at the Charleston brig and then develop computer software capable of identifying and verifying individual inmates based on their biometric characteristics. The software also had to be easy enough to operate so that corrections specialists with limited prior training or experience on computers could understand how to use it.


All biometric methods — iris, facial, retinal, finger and hand geometry, voice and fingerprint — were tested over a three-year period. All had been developed, tested, and used in other settings, mostly by commercial firms. And all were found to have advantages and disadvantages at the Charleston brig. Facial recognition produced too many false positives on prisoners. Although biometric methods do not have to work every time to be effective, corrections specialists had to visually identify the prisoners too often, thus slowing the process. Iris recognition was the most accurate method tested at the Charleston brig, but it was similarly judged too slow to work effectively in a jail setting. Voice recognition proved to be the least accurate method tested.


In the end, the fingerprint recognition method, now used in conjunction with hand geometry, was judged to work best at the Charleston brig. It provided the most accurate and reliable matches at about one-third the cost of iris, facial and retinal methods. The fingerprint method also moved prisoners through the gates faster than the others. That’s a prime consideration when, for example, corrections specialists are moving 50 or more prisoners at once from housing or work areas to the galley at mealtime. Fingerprint readers were also easier to use and more durable than other readers.


In the next phase, the manual dry-erase board and paper system was replaced with a computerized tracking system in which a server contained all data on inmate movements. Brig staff could access the data from each housing unit, the control center and the enrollment area. Biometric scanners were then added to further verify the location of prisoners.


As the system now works, the computer finds a biometric match, identifies the individual prisoner and confirms that he or she is authorized to go from one part of the brig to another. The computer also sends a message to the next security post on the prisoner’s authorized path that the prisoner is on his or her way. No escort or paper record is necessary because the computer records all prisoner movement between security posts at different parts of the brig. If a prisoner fails to show up within a specified time, usually five minutes, an alarm is sounded and the staff are alerted that a prisoner is out of place.


Evaluating the Project


Outside experts were engaged to help NIJ and the U.S. Navy evaluate how well the computer tracking and biometrics systems worked. Initially, evaluators conducted surveys of brig staff taken before the biometric system was fully in place. The surveys showed that the corrections specialists and other brig officials thought the existing system for tracking prisoner movements worked fine most of the time.


To test that assumption, evaluators asked brig officials to “grab” and hold a prisoner who was authorized to move from one part of the brig to another after that prisoner had passed through the first security post. By so doing, they deliberately caused a prisoner to be late and out of place, thus creating a security breach. The results showed that the Charleston brig’s manual system did not work as well as its staff had thought.


Under the manual system of tracking inmate movements, the corrections specialists failed to note a prisoner’s non-arrival in all 12 test grabs. Under the manual system, it took corrections specialists an average of 43 minutes to notice an out-of-place prisoner. In half the cases, more than one hour passed before the corrections specialists realized the situation. Once the computer tracking system was introduced, however, the average time it took for staff to notice a non-arriving inmate dropped to 17 minutes. In only one of 10 cases did more than one hour pass.


At the same time, the computer tracking system improved the efficiency of corrections specialists and other brig officials. Most corrections specialists learned the new system quickly, which, when mastered, calls for less reliance on their memory of individual prisoners and provides automatic warnings when prisoners are deemed out of place. The system frees corrections specialists from handling paper passes, allowing them to spend more time actually watching prisoners in their area. And that translates into improved staff efficiency. So, too, do the fewer outgoing telephone and radio calls made by the corrections specialists to ensure that a prisoner has actually arrived at the next post on time. Once again, that means more time for staff to spend actually watching prisoners.


NIJ and project staff plan to take what has been learned at the Charleston naval brig and apply it to a larger, civilian prison. The goal is to develop the technology, software, and methods to use biometrics in any prison or jail in the United States.


Challenges lie ahead: Technological developments continue to change the relative merits of the different recognition and verification methods. A civilian prison represents a riskier, higher-use setting than a military jail for biometrics to work, so a different set of criteria will need to be developed to evaluate which systems work best. For example, equipment durability may be more important because of higher volumes of use and because of the increased potential for deliberate vandalism by inmates likely to damage equipment designed to track their movements.


The Charleston brig test is not yet completed—final evaluations of the full biometric and computer system remain to be finished and analyzed. Yet the project staff is optimistic that with further testing and analysis, biometrics technology can be used successfully in U.S. prisons and jails to identify and track inmates. n


Christopher A. Miles is a senior program manager for research and technology at NIJ. Jeffrey P. Cohn is a freelance writer/reporter. Their article was originally printed in NIJ Journal.