Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-12-01
2003-01-07
Amsbury, Wayne (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
Reexamination Certificate
active
06505193
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates in general to identification of physical characteristics of a human being, and particularly, the present invention relates to a system and method of fast biometric database searching using digital certificates. More particularly, the present invention relates to using iris recognition and digital certificates to conduct fast biometric database searching of relatively large databases for the identification of a subject.
BACKGROUND OF THE INVENTION
Various techniques are used for uniquely authenticating a remote server provider or computing platform. For example, encrypted passwords, account information, and Personal Identification Numbers (PIN) have been used as tools to authenticate a computing platform and to authorize a transaction originating at the computing platform.
More recently, certificates have been used to authenticate a computing platform as being an authorized platform for a particular application. Perhaps the best known public-key certificate format is X.509, which is an identity-based design. That is, it ties a public key to a computing platform, such as a distinguished name. The distinguished name is a unique string supposedly associated with a computing platform or other object. A certifying authority assigns each certificate and has the task of ensuring that each certificate is unique and that the certificate is an accurate and true label for the computing platform.
While the above authentication techniques provide a mechanism for authenticating a computing platform as being an authorized platform for a given application, these technologies do not provide for the identification of the person at the computing platform who is using the computing platform to seek services from the application server. They simply provide a token in place of identity.
Various technologies are used for uniquely identifying a person in accordance with an examination of particular attributes of either the person's interior or exterior eye. One of these technologies involves the visual examination of the particular attributes of the exterior of the iris of at least one of the person's eyes. The iris of the human eye has random patterns of striations, ciliary processes, crypts, rings, furrows and other features which have been shown capable of generating highly unique biometric templates for personal identification. In this regard, reference is made to U.S. Pat. No. 4,641,349, “Iris Recognition System”, issued to Flom et al., and U.S. Pat. No. 5,291,560, “Biometric Personal Identification System Based on Iris Analysis”, issued to Daugman. As made clear by these patents, the visible texture of a person's iris can be used to distinguish one person from another with great accuracy. Thus, iris recognition can be used for such purposes as controlling access to a secure facility or a bank automatic teller machine, for example. An iris recognition system involves the use of an imager to video image the iris of each person attempting access, and image processing means for comparing this iris video image with a reference iris image on file in a database.
Iris identification systems have been developed that are capable of collecting images of the iris and processing them to produce biometric templates. These templates may be used to identify human individual irises with extremely low error rates, on the order of 1 in 10
6
.
Iris recognition is widely recognized as the most powerful technology available for biometric identification of humans. Most biometric technologies are useful only for verification, implying 1:1 matching between a live biometric measurement and a single stored template. However, some biometric technologies, such as the iris recognition technology developed by IriScan®, are capable of identification, in which the live biometric is matched against N entries in a database to identify the individual without the need to provide a presumed identity or biometric token. This 1:N match can be performed for very large N due to the extremely low single-match false accept and false reject rates which characterize some biometric technologies, such as the IriScan® iris recognition technology. However, challenges are posed by 1:N matching of biometric templates when N is very large, and match times can quickly become unacceptable.
The template matching process, when viewed as a database search operation, has some very demanding requirements. First, there is no “key” that can be used to arrange the templates in the database in some advantageous way to facilitate searching. There is no way to avoid matching the unknown template against each and every database entry until the correct match is found, or the search is exhausted. Also, each match is typically performed at a number of different relative rotational alignments of the two templates. So each pair of templates really requires as many as 21 separate bit comparisons, accounting for ±28 degrees of relative rotation. For example, a typical single 300 MHz Pentium-based PC can match templates at a rate of approximately 40,000 matches in about a second or two.
This may be entirely adequate for most physical access control installations, in which the database has less than about 40,000 entries and a match time of a second or two, or longer, is acceptable. However, for large networked systems where a single access server has to process multiple simultaneous matching requests, over a secure network with a database of millions of enrollees, match times can quickly become unacceptable. Typically, an exhaustive search is performed for each unknown template, and consequently, the match speed varies directly with the processing power and inversely with the number of simultaneous requests.
There are a number of conventional strategies for increasing the match speed that yield low-confidence matching. One is “filtering”, in which information like sex, birth date, etc. is used to pre-classify the unknown person to be identified. This filtering technique is used, for example, with the less accurate finger-print identification techniques. The database is partitioned according to these pre-classifications, and when the person is enrolled the biometric template is placed in the correct partition. When the person seeks to be identified, they must provide information (age, sex, etc.) to identify the database partition containing their template. However, if the pre-classification is done incorrectly, either at enrollment or later at identification, the search will be conducted over the wrong segment of the database and the person will never be found. As a result this strategy produces high false reject rates and completely unreliable results.
The matching speed can be increased through the use of multiple processors in parallel. The match speed achieved is linearly related to the speed and/or number of processors. Although this is a viable approach, it is not cost-effective. Greater gains might be achieved by designing and building specialized processors that can perform the simple XOR logic operations at very high speed. However, this speed advantage is limited by the rates at which data can be moved into and out of the processor.
Although the art of biometric recognition systems is well developed, there remain some problems inherent in this technology, particularly with biometric systems and methods characterized by 1:N matching of biometric templates when N is very large. A system and method for achieving fast, accurate, cost-effective identification in these applications is needed. Therefore, a need exists for a recognition system and fast search methodology that overcomes the drawbacks of the prior art.
SUMMARY OF THE INVENTION
The present invention is directed to a system and method for conducting fast biometric database searches using iris recognition and digital certificates. The present invention provides for the authentication of a computing platform based on digital certificates attached thereto and also for the relatively fast identification of a person at the computing pla
Cambier James L.
Musgrave Clyde
Amsbury Wayne
Iridian Technologies, Inc.
Nguyen Cam-Linh T.
Woodcock & Washburn LLP
LandOfFree
System and method of fast biometric database searching using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method of fast biometric database searching using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method of fast biometric database searching using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3065658