Image analysis – Applications – Personnel identification
Reexamination Certificate
1999-06-01
2001-05-08
Mehta, Bhavesh (Department: 2621)
Image analysis
Applications
Personnel identification
C382S124000, C382S117000
Reexamination Certificate
active
06229906
ABSTRACT:
FIELD OF THE INVENTION
The present invention teaches a technique of identifying a user through sequential input of human body part information. More specifically, the present invention teaches a technique of recognizing a specified sequence of user's fingerprints as a key to identification of the person.
BACKGROUND AND SUMMARY
Biometrics is a study of using a physical trait in a human body part for verifying identity. Various techniques in the art have been suggested including recognition of fingerprints, recognition of retinal scans, palm prints, DNA traces, and others. Each of these human body part identification techniques has its own host of limitations.
Fingerprint identification is extremely reliable, when used properly. However, proper identification of a fingerprint requires massive amounts of computer power. A reasonably-sized computer system might take two to three seconds to identify a fingerprint with a reasonable degree of certainty. Moreover, smudges and other imperfect entry of fingerprint information can often leave traces which hinder accurate verification of the fingerprints.
Users rapidly lose patience with any system that gives them a false rejection. Hence, it is often desirable to leave the logic “wide open”; that is to allow recognition of a fingerprint with only a minimal degree of matching relative to the stored original. This tradeoff makes it less likely that a proper fingerprint will be erroneously rejected. However, it makes it more likely that an improper fingerprint will be erroneously recognized as being proper. The processing necessary to make a rough identification of such a fingerprint is less than that necessary to make a detailed identification, allowing a quicker recognition.
Some have suggested that a fingerprint detection system, especially a loose-logic system like the above-described, could be bypassed by taking a photocopy of a fingerprint of an appropriate type. Others have grotesquely suggested cutting off the user's finger or keeping the user as a captive in order to use their fingerprint.
The same problems do not exist in a personal identification system based on confidential personal identification numbers (PINs). Conventional ways to form a PIN include using a series of numbers, or a series of alphabetic letters, or a combination of letters and numbers. A secret PIN is known only by the user. Detection of the entry of the correct PIN is relatively simple. The system cannot be fooled by a fake PIN. However, various techniques are available to obtain a PIN surreptitiously. The user can be forced to tell the PIN number to another. A PIN number can be observed, e.g. by watching the user enter it. Since any user can enter the PIN number, its security depends on secrecy.
It is an object of the present invention to adapt the systems described above. Specifically, the present inventors devised a system which uses some features in body parts for identification and in addition implements a secret code formed by a sequence of body parts to further enhance the identification. The inventors believe that use of the unique coding features in a user body part verification, in conjunction with the secrecy of personal identification numbers combines some of the best features of both identification systems. The logic can be left wide open (i.e., a loose-logic) to allow quick recognition of body parts, e.g., fingerprints. The present invention, however, goes against the established teaching in a biometric identification system to further make a positive identification by requiring that secret information be entered as a part of the biometric information, e.g., a secret sequence formed by fingerprints.
The inventors recognized the distinct advantages of a biometric identification systems by employing some bio-components in human's body parts to obtain certain information usually unique only to a particular person. The body parts used in the present invention are readily available to the user at anytime and anywhere. The present invention uses fingerprints in preferred embodiments disclosed herein to exemplify the concept of the invention.
The inventors recognized that multiple bio-components of a person can be used for redundancy and enhancing the possibility of a positive identification. Multiple fingerprints of a person to be identified, for example, are used in the present invention to obtain multiple matches in comparison of the entered fingerprints with the fingerprints of a user stored in a database.
Importantly, the inventors recognized the limitations in various biometric identification systems including system processing speed, identification ambiguity, and security. In particular, the inventors recognized the identification ambiguity and security issues in fingerprint identification. The present invention implements a novel personal identification coding system with a sequence code of bio-components of human body parts. Use of sequence codes formed by body parts in accordance with the present invention significantly increases the possibility of a positive identification and provides additional security against unauthorized use of body parts. Moreover, sequence codes of the present invention allows a coarse recognition in body part recognition without compromising the system performance. This reduces the computation that is required to positively recognize body features.
In the preferred embodiments of the present invention, multiple fingerprints of a person (e.g., three fingerprints of a person) are entered and processed to identify a particular user in the system database. A confidential sequence code constructed by multiple fingerprints is used to make a final positive identification of the user. For example, a sequence is formed by five fingerprints selected from three different fingers of a user. Examples of various encoding methods are described herein.
The inventors further recognized the importance of tamper-resistance in such identification systems. In order to be positively recognized by the system of the present invention, an intruder's bio-components in body parts, e.g., fingerprints, must be very similar to a user's information stored in the system database. Even if the intruder obtains a false fingerprint or a user's real fingerprints, or somehow fools the identification system in fingerprint matching, the intruder still does not know the secret information, e.g., the proper personal identification code. Moreover, even if the intruder were to observe a user entering the secret code, the intruder still would not be able to re-enter that code, because the intruder needs the body part for a biometric matching (e.g., a matching of fingerprints).
In addition, the inventors recognized the technical problems and challenges in biometric identification systems which compare entered bio-components of human body parts (e.g., fingerprints) with the corresponding data of a user in the system database. For example, large amount of data is to be processed in such a system identifying fingerprints since images of entered fingerprints are compared with stored images in the system database and specially in a large database common in many commercial applications. Electronic processing techniques often take a long time to correlate the input fingerprint images with the stored images in the database due to the limitations of the microprocessors and the serial processing nature.
One aspect of the present invention defines using the parallel processing capability as well as the fast processing speed of an optical correlator to significantly increase the processing speed of the identification system disclosed herein. In particular, multiplexed volume holographic storage media are used in the optical correlator incorporated in the present invention for fast data accessing and image processing.
The present invention can also be applied to electronic processing systems that use Digital Signal Processing (DSP) algorithms for matching body features. One unique aspect of such a system is that it can leave the logic wide open
Psaltis Demetri
Pu Allen
California Institute of Technology
Fish & Richardson P.C.
Mehta Bhavesh
LandOfFree
Biometric sequence codes does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Biometric sequence codes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Biometric sequence codes will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2502590