Hand and finger geometry recognition is the process of identifying an individual through the unique "geometry" (shape, thickness, length, width, etc.) of that individual's hand or fingers. Hand geometry recognition has been employed since the early 1980's and is among the most widely-used biometric technologies for controlling access to important assets. It is easy to install and use, and is appropriate for use in any location requiring highly-accurate, non-intrusive biometric security. For example, it is currently used in numerous workplaces, day care facilities, hospitals, universities, airports, and power plants.

A newer option within hand geometry recognition technology is finger geometry recognition (not to be confused with fingerprint recognition). Finger geometry recognition relies on the same scanning methods and technologies as does hand geometry recognition, but the scanner only scans two of the user's fingers, as opposed to his entire hand. Finger geometry recognition has been in commercial use since the mid 1990's and is mainly used in time and attendance applications (i.e., to track when individuals have entered and exited a location). To date, the only large-scale commercial use of two finger geometry for controlling access is at Disney World, where season pass holders use the geometry of their index and middle finger to gain access to the facilities.

To use a hand or finger geometry unit, an individual presents his or her hand or fingers to the biometric unit for "scanning." The scanner consists of a Charged Coupled Device (CCD), which is essentially a high resolution digital camera; a reflective platen on which the hand is placed; and a mirror or mirrors that help capture different angles of the hand or fingers. The camera "scans" individual geometric characteristics of the hand or fingers by taking multiple images while the user's hand rests on the reflective platen. The camera also captures "depth," or three-dimensional information, through light reflected from the mirrors and the reflective platen. This live image is then compared to a "template" that was previously established for that individual when they were "enrolled" in the system (see below for a full discussion of the enrollment process). If the live scan of the individual matches the stored template, the individual is "verified," and is given access to that asset (see Attributes and Features for a discussion of verification). Typically, verification takes about 2 seconds. In access control applications, the scanner is usually connected to some sort of electronic lock, which unlocks the door, turnstile, or other entry barrier when the user is verified. The user can then proceed through the entrance. In time and attendance applications, the time that an individual checks in and out of a location is stored for later use.

As discussed above, hand and finger geometry recognition systems can be used in several different types of applications, including access control and time and attendance tracking. While time and attendance tracking can be used for security, it is primarily used for operations and payroll purposes (i.e., clocking in and clocking out). In contrast, access control applications are more likely to be security-related. Biometric systems are widely used for access control, and can be used on various types of assets, including entryways, computers, vehicles, etc. However, because of their size, hand/finger recognition systems are primarily used in entryway access control applications (see discussions of size considerations below). Therefore, this document will focus on hand/finger geometry recognition for entry access control.

Registration (termed "enrollment") in a hand or finger geometry recognition system is the process by which a record or template of the enrollee's hand or finger geometry is established. As described above, live scans are compared to this template when a user attempts to access an asset. If the live scan matches the template, the user is given access to that asset.

Enrollment generally takes about 30 seconds per user. To enroll in either type of system, the enrollee's hand is placed palm down on the reader's surface. The placement of the hand is aided by pins that serve as guides to correctly orient the hand and fingers for the camera. The camera then takes an image of the hand/fingers and stores it. The unit's internal processor and software convert the image of the hand/fingers to a mathematical representation, which is then compressed by an algorithm and stored as the user's template. The template may reside in the biometric unit's internal memory, in a centralized database of users, or on other media, such as on a smart card or a hard disk.

One potential problem with these types of systems is the slight variability (even with the use of guides) in the placement of a user's hand or fingers onto the biometric unit. The resulting scan will reflect this slight variability and might impact the resulting "access decision," resulting in a false rejection when a user attempts to access an asset. Some systems may be able to compensate for this variability by actually incorporating it into the template. For example, enrollment in Recognition Systems Inc.'s hand scanning system requires that the enrollee's hand be placed on the platen and removed from the surface three separate times to capture three separate images. The three images are averaged and combined to form a single template that includes the slight variations between the three measurements. Thus, when the user scans his hand when attempting to access an asset, the scan is compared to a template that has already included some variation. If there is enough of a match between the template and the scan, the match is accepted.

The hand and finger geometry recognition systems currently on the market can typically take over 90 measurements of the hand or fingers. The camera "scans" features including, but not limited to: finger length and thickness; the distance between joints; the length, width, and thickness of the hand; the surface area of the hand; and the overall bone structure of the hand and fingers. Each measurement is mapped as a point on a three-dimensional grid that forms a mathematical representation of the hand/fingers. As described above, the mathematical model of these scanned features is compressed using an algorithm and is stored as a template.

In order to produce as accurate and precise a template as possible, the camera/scanner is programmed to record biotraits that should not change from scan to scan. It does not record superficial surface details - such as lines, scars, dirt, fingerprints, and fingernails- which are more likely to change from one scan to another. Because the system records and compares biotraits that should not change from one scan to the next, the chances of false rejection due to variations between the scan and the stored template should be minimized.

Current hand and finger recognition systems use a one-to-one "verification" process to identify individuals. In a verification process, the individual first identifies themselves as a specific person, usually by typing a password or code into a keypad or by presenting a proximity card or smartcard, etc. Once the individual has informed the system as to who he is supposed to be, he then submits his hand/fingers to the biometric reader to be scanned. The software in the system compares this scan with the template that the user has identified as his. For example, if the user enters a password or code on the keyboard that identifies him as John Smith, the scanner will attempt to match ONLY John Smith's template with the bioscan. If the scan and the template match, the individual will be granted access.

Because similarities between human hands are not uncommon, current hand/finger recognition systems cannot operate in a one-to-many comparison "identification" mode, whereby the user's biotrait is scanned and the software attempts a match versus all of the profiles stored in its system. This will limit their use in many applications where identification rather than authentication or verification is desired or required. For a full discussion of one-to-one verification processes versus one-to-many identification processes, please refer to the Authentication vs. Identification section of the Biometric Security Systems Security Product Guide.

Since hand and finger geometry recognition systems perform a one-to-one verification comparison, false rejection rates are among the lowest of any biometric technology. Sandia National Laboratories has documented Recognition Systems' hand geometry scanners as having a false rejection rate of 0.1 percent (1 false rejection in 1000 scans). According to the vendor, a false rejection rate of approximately 0.3 percent (3 in 1000) can be expected for the finger geometry scanner currently on the market if the baseline settings are used. The accuracy of each system can be adjusted, but there are some drawbacks to doing so (for example, changing the system sensitivity to reduce false rejections may increase false acceptances). For more information on the relationship between high sensitivity and false rejection rates, see the Accuracy section of the Biometric Security Systems Security Product Guide.

Hand and finger geometry recognition systems are fairly resistant to fraud. Because these systems use one-to-one verification comparisons when attempting to match to a template, a would-be intruder would have to be able to submit an authorized person's identification information, and then submit a model of the user's hands or fingers to the reader, in order to defeat the system. This would be both difficult and time consuming.

Hand and finger geometry recognition systems are easy to use, and with minimal training users can to learn the correct placement of their hand/fingers onto the platen.

One potential drawback of any biometric system is users' attitudes towards them. However, attitudes towards hand and finger geometry scanners are usually positive. People generally do not regard the placing of one's hand on a reader as intrusive. This contrasts with some other forms of biometric identification, which are considered more intrusive and may have a stigma associated with them. For example, a retinal scanner sends an infrared light into the eye, which some people dislike; and some individuals associate the fingerprint scanner with criminal activity.

Hand and finger scanning units require minimal data storage capabilities because the templates are typically small. For example, the template size for Recognition Systems Inc.'s hand scanner is approximately 9 bytes, while the template for Biomet Partners' finger scanner is 20 bytes. This data storage requirement is several orders of magnitude smaller than that required for most other biometric technologies. For example, the template for a fingerprint scan ranges from 250-1,000 bytes, while biometric voice-scan templates are generally 1,500-3,000 bytes.

The small template size for hand and finger geometry recognition facilitates storing a large number of templates. This minimizes the storage space required for centralized systems and allows entire user databases to be stored in stand-alone units. For example, a single hand geometry scanner unit from Recognition Systems, Inc. can store up to 32,512 templates. Its small size also permits the template to be stored on nearly all card-based media.

Unlike other biometric technologies, which function best in controlled environments with minimal temperature variances and a high degree of cleanliness, hand and finger scanners function fairly well under a wide range of temperatures and environments. Because the scanner does not examine the minute details of the hand and fingers, but captures more of their overall shape, it works well in harsher environments where cleanliness is not always possible, and in cases where minor changes, such as scratches, cuts, or dirt, can affect the user's hands. For these reasons, these types of scanners are widely used to control access to facilities where manual labor is performed.

These units also have minimal maintenance requirements. Unlike other types of biometric devices that require physical contact with the surface of the unit and which must be cleaned often to keep their surfaces free of smudges, dirt, and debris, the hand and finger geometry scanners function well with only a moderate level of maintenance. Hand and finger scanning units can generally be cleaned once every two or three weeks. Cleaning can be done with glass cleaner and a soft cloth.

As with any biometric trait, change in the biometric trait can result in false rejections. Since injuries to the hand and fingers are fairly common, an injury causing swelling, or injuries requiring a cast, splint, or bandage, may affect a user's scan, and may result in a false rejection. To compensate for this possibility, most hand geometry scanning units will allow for registration of the opposite hand, and most finger geometry scanning units will accept readings from the fingers of either hand. However, other backup systems, such as a separate backup biometric unit, a mechanical lock, or an intercom system, would need to be implemented in the event of injuries to both hands.

There may be size issues for these devices. Although they are relatively small in size, hand and finger scanners tend to be larger than most other biometric scanners. Therefore, they may be precluded from designs that require smaller profiles. Since the hand scanner must be large enough to accommodate the human hand, it will not be as able to capitalize on new advances in technology that would make the unit smaller in size.

The image capturing and verification software and hardware for the hand and finger geometry recognition units that are currently on the market allow these systems to exist as stand-alone units or to be integrated within a multi-unit system (i.e., units at multiple access points are integrated into one system). Applications that require a large number of access points and users can be administered at a central location. Having a centralized database of users eliminates the need to store template information on each individual biometric unit and eliminates the need for individuals to register on each device. For further information on stand-alone and centrally-administered units, see the discussion of the Host Computer in the Biometric Security Systems Security Product Guide.

The hand and finger geometry recognition systems currently on the market are easily integrated with existing card-based systems. They can also be easily combined with identification-based systems. For example, the hand and finger geometry recognition system may serve as the primary access control and the identification system (such as a fingerprint reader) may serve as a backup. In this example, if a user has a bandage or a cast on his hand, and either cannot use the hand/finger geometry recognition system or is rejected by it because of the cast or bandage, he could use the fingerprint reader as an alternative to gain access to the protected asset.

Stand-alone units are available for both hand and finger geometry scanners. These units generally cost from $1,000-$3,000, depending on which options are chosen and on the quantity ordered. Typically, stand-alone units include all of the hardware and software to operate a single, independent unit at one control point (i.e., a doorway or some other entryway). The lower-end modules typically include a numeric keypad and an LCD display, while upgraded models might include a card-reader interface, additional memory, and a more durable metal casing unit. Costs for the unit do not include installation. In a typical access control application, the unit must also be integrated with some sort of locking system, whether it is an electronic door lock or some other type of lock. Adding any specialized locks to the entryway and integrating them with the biometric unit will also add additional costs to the system. Finally, while stand-alone units will have low configuration and overhead costs, they may be limited in their ability to support future expansion.

Integrating the hand or two finger geometry reader with existing security systems and customizing these components (i.e., networking multiple units at different locations) will add significant costs to any biometric security system. Since options and existing conditions vary widely, vendors should be contacted for specific pricing.