A survey of biometric technology based on hand shape
Introduction
Automatic human identification has become an important issue in today's global information society. Due to increasing security concerns, a large number of systems currently require positive identification before allowing an individual to use their services. During the last decade, there has been a steady research effort1 towards providing user-friendly and reliable methodologies for access to facilities, resources, services or computer systems. Biometric systems are already employed in domains that require some sort of user verification. It is generally accepted that physical traits like iris, fingerprints and, as more recently debated, hand shape and palmprints can uniquely define each member of a large population which makes them suitable for large-scale identification (establishing a subject's identity) [10], [22], [50]. On the other hand, in many small-population applications, because of privacy or limited resources, we only need to verify a person (confirm or deny the person's claimed identity). In these situations, one can also use behavioral traits2 which have less discriminating power such as voice, face, signature and human–computer interaction (HCI) derived patterns.
Comprehensive reviews have been recently published on fingerprint [31], palmprint [50], face [49] and behavioral HCI [46] biometric technologies. In this paper, we survey the state of the art in hand shape-based biometric technology and complement the information provided in [10], [36]. We also describe some of the commercially deployed systems and analyze the practical issues concerning enrollment, training and performance evaluation to consider when designing such systems.
Hand shape biometrics is the ensemble of techniques employed in establishing the identity of a person using person's hand silhouette and/or geometric features (e.g. finger lengths, widths, areas, ratios, etc) derived from it. In the biometric literature, hand shape systems usually include all systems using information extracted from a hand silhouette, while hand geometry refers to only those systems which use sparse geometric features. A typical hand shape biometric system uses a camera or scanner-based device to capture the hand image of a person and compares this against the information stored in a database to establish identity. Besides person identification, hand imaging has also been used for deriving statistical models of biological shapes [14] and for guiding gesture-based HCI tasks [43].
As often noted in the literature, hand shape biometrics is attractive due to the following reasons:
- 1.
Hand shape can be captured in a relatively user convenient, non-intrusive manner by using inexpensive sensors [2], [25], [30].
- 2.
Extracting the hand shape information requires only low resolution images and the user templates can be efficiently stored (nine-byte templates are used by some commercial hand recognition systems [35]).
- 3.
This biometric modality is more acceptable to the public mainly because it lacks criminal connotation [16], [24].
- 4.
Additional biometric features such as palmprints and fingerprints can be easily integrated to an existing hand shape-based biometric system [10], [25], [30].
Section snippets
Operation of a hand shape-based biometric system
A hand shape-based biometric system operates according to Fig. 1. In the enrollment stage, hand shape data are acquired from the registered users, feature sets are extracted from the acquired data, and one or multiple templates per individual are computed and stored in a database. In the deployment stage, one snapshot of the user's hand is captured, a feature set is computed and then compared to the templates in the database. Based on the comparison result, the claimed identity is accepted or
System taxonomies
A biometric taxonomy is usually based on partitions which classify the usage of biometrics within a given application. The following taxonomies have been mentioned in the literature (biometric systems are considered to perform best when employed in the first five conditions of the left column) [32], [39]:
- •
Cooperative vs. non-cooperative user. (Are users willingly participating?)
- •
Overt vs. covert biometric system. (Are users knowingly participating?)
- •
Habituated vs. non-habituated user. (Are users
Deployment history
Hand geometry appears to be the oldest automatically employed biometric modality as hand geometry-based systems have been commercialized for almost four decades [27], [36], [52]. These systems accounted for 4.7% (141 million US dollars) of the world revenues generated by all biometric modalities and were the fourth leading biometric technology in 2007 [18]. The technology has evolved from stand-alone electro-mechanical devices in the early 1970's to today's electronic scanners attached to
Hand shape uniqueness
The issue of hand shape uniqueness within a large population is currently somewhat controversial. Some researchers and designers of commercial biometric systems consider hand shape to have a medium-to-high discrimination power [30], [32]. On the other hand, the authors of some recent research systems [10], [47], [48] have shown high verification/identification rates which are comparable to fingerprint-based systems. However, their datasets appear to be collected offline and have, arguably,
Future directions
Hand shaped-based authentication systems have been widely used for applications in access control, attendance tracking and personal identity verification for almost 40 years. However, this is still an active area of research which focuses on:
- (i)
Designing an unconstrained hand image acquisition setup, in which no guiding pins or even a platform are needed. That can be achieved by using different illumination [29], multi-modal sensors (e.g. combined 2-D and 3-D range [41]), sophisticate silhouette
About the Author—NICOLAE DUTA received the B.S. degree in applied mathematics from the University of Bucharest (Romania) in 1991, the D.E.A. degree in statistics from the University of Paris-Sud (France) in 1992, the M.S. degree in computer science from the University of Iowa in 1996 and the Ph.D. degree in computer science and engineering from Michigan State University in 2000. He is currently a scientist in the Natural Language Understanding group at Nuance Communications, Burlington, MA.
References (52)
- et al.
Biometric verification/identification based on hands natural layout
Image and Vision Computing
(2008) - et al.
Personal authentication using hand images
Pattern Recognition Letters
(2006) - et al.
Finger surface as a biometric identifier
Computer Vision and Image Understanding
(2005) - et al.
Model-guided deformable hand shape recognition without positioning aids
Pattern Recognition
(2005) - et al.
Hand biometrics
Image and Vision Computing
(2006) - et al.
Peg-free hand shape verification using high order Zernike moments
- ANSI INCITS 396-2005, Hand geometry interchange format, available at:...
- Biomet Partners Inc., FingerFoto™, available at:...
- et al.
Hand recognition using geometric classifiers
- Business wire, IR recognition systems biometric handreaders support BioAPI, 2003, available at:...
Personal verification and identification using hand geometry
ECTI Transactions on Computer and Information Technology
A comparative analysis of global hand appearance-based person recognition
Journal of Electronic Imaging
Sheep, goats, lambs and woves: a statistical analysis of speaker performance in the NIST 1998 speaker recognition evaluation
On the error-reject trade-off in biometric verification systems
IEEE Transactions on Pattern Analysis and Machine Intelligence
HANDS: A Pattern-Theoretic Study of Biological Shapes
Models of large population recognition performance
Deformable matching of hand shapes for user verification
A prototype hand geometry-based verification system
Introduction to biometrics
Cited by (150)
A comprehensive survey on the biometric recognition systems based on physiological and behavioral modalities
2020, Expert Systems with ApplicationsGait Recognition Algorithm of Coal Mine Personnel Based on LoRa
2023, Applied Sciences (Switzerland)Modeling and analysis of upper limb kinematics in sign language recognition
2023, Research SquareBiometric security performance: Analysis methods and tools for evaluation and assessment
2023, Artificial Intelligence for Biometrics and Cybersecurity: Technology and applicationsPalmEcho: Multimodal Authentication for Smartwatch via Beating Gestures
2023, Annual IEEE Communications Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks workshops
About the Author—NICOLAE DUTA received the B.S. degree in applied mathematics from the University of Bucharest (Romania) in 1991, the D.E.A. degree in statistics from the University of Paris-Sud (France) in 1992, the M.S. degree in computer science from the University of Iowa in 1996 and the Ph.D. degree in computer science and engineering from Michigan State University in 2000. He is currently a scientist in the Natural Language Understanding group at Nuance Communications, Burlington, MA. From 2000 to 2005 he was a scientist in the Speech and Language Processing department at BBN Technologies, Cambridge, MA. He also held temporary research positions at INRIA-Rocquecourt (France) in 1993 and Siemens Corporate Research (Princeton, NJ) from 1997 to 1999. He is a member of IEEE and his current research interests include computer vision, pattern recognition, language understanding, automatic translation, machine and biological learning.