Image analysis – Applications – Personnel identification
Reexamination Certificate
1999-06-10
2002-06-25
Mehta, Bhavesh (Department: 2621)
Image analysis
Applications
Personnel identification
Reexamination Certificate
active
06411728
ABSTRACT:
FIELD
This invention relates generally to a method and apparatus for identifying an individual from an examination of their natural physical characteristics, specifically, an examination of finger pores and associated macrofeatures.
BACKGROUND
Credit card and check fraud cost the financial industry billions of dollars every year. The reason is that these financial instruments are used only with a token such as a credit card which acts only to identify the individual's financial accounts Often, only a credit card number stands in the way of the fraud perpetrator.
Although PIN numbers and social security numbers are also used to limit fraud, as is well known, PIN numbers and social security numbers are easily obtained by the fraud perpetrator through various surreptitious methods. In order to prevent these types of fraud, many have proposed the use of biometrics (the use of unique physical characteristics of individuals), to identify an individual, in conjunction with verifying or identifying the individual through a credit card, ATM card, account number, or the like.
The methods for identification of individuals using fingerprints based upon examination of ridge based data are well known and are discussed in references such as patents to Costello (U.S. Pat. No. 5,321,765), Schiller (U.S. Pat. No. 4,752,966) and Kim (U.S. Pat. No. 5,105,467). As described in Schiller, biometric identification of individuals is associated with two types of errors. These errors are widely known in the industry as False Reject (Type I) and False Accept (Type II) errors. False accept errors occur when there are enough similarities between the fingerprints of two individuals that one is mistaken for the other. A false reject error occurs for a variety of reasons, and refers to when an individual is not identified even though the individual is an authorized user registered with the system.
However, identification of individuals through the examination of ridge based data poses problems. One drawback of current fingerprint identification systems is that they have non-zero false accept and false reject rates. In a large scale system capable of handling the approximately 110 million U.S. credit card holders, a biometric identification system with even modest false accept and false reject rates may end up creating more in losses due to mistaken identity than it prevents by eliminating criminal fraud. Improving the false accept rate by only a few percentage points for 110 million users amounts to preventing millions of dollars in losses.
Others have suggested that the underlying cause of false accepts and false rejects is that the amount of data from a fingerprint is too limited for it to be used in a biometric identification system involving large numbers of users. The typical number of minutia points available in a fingerprint is reported to be about 30 to 40. Other biometric technologies have emerged to fill this void. Use of iris biometric data has been touted to increase the number of available data points to about 266 features.
Of the above-mentioned references, all uniformly disregard the presence of pores (also known as sweat pores or finger pores) on fingerprints. Pores are naturally occurring physical characteristics of the skin which exist upon the ridges of the fingerprint. Some, including Schiller, consider pores to be minor holes which should be disregarded in what is otherwise a continuous ridge line structure. A typical finger contains about 50 to 300 pores which can be used as additional sources of features for uniquely identifying an individual.
A reference which recognizes that pore data from fingerprints can be used as a source of information for producing either a failed or successful identification of individuals is a paper entitled
Automated System for Fingerprint Authentication.
(Stosz, et al., SPIE, Vol. 2277, p. 210-223). This reference describes a process for image processing of raw biometric images, including gray scale to binary conversion, skeleton graph processing, thinning, and healing, which are well known in the art.
Stosz also describes a multilevel verification process wherein pore locations and minutia data are used separately to confirm or cross-check the identity of individuals. That is, initially pore locations are matched against pore locations and a correlation score is obtained from the pore matching which results in either a successful or failed identification result. Next, assuming the pore match indicated a successful identification result, minutia points are independently matched to minutia points to verify the identification established by pore matching.
A problem with matching pores by location alone occurs when a finger with a high pore density is presented. Simple tests show that a 50% correlation score between two large sets of randomly located pores is possible when used with a location proximity threshold set to about the length of the maximum pore size. These results indicate that a system which matches pores by location alone will have difficulty discriminating between two fingerprints which both have large numbers of pores.
As suggested by Stosz, a high-resolution scanner of at least 800 DPI or greater is required to accurately resolve pores. However, high-resolution scanners have several disadvantages when compared to low-resolution scanners (500 DPI or less). These disadvantages include capturing only a partial fingerprint image in that a similarly sized sensor at a high resolution resolves a smaller area of the finger being imaged than a low-resolution scanner. This is similar to when a microscope lens is adjusted to a higher magnification. To get around this problem, a larger CCD or CMOS image array can be used, however, this results in a significantly higher cost. Other methods can be incorporated to solve this problem such as a movable scanner head. However, all are cumbersome, expensive, unreliable, or time consuming to use. Additionally, the FBI has adopted a 500 DPI standard for its fingerprint images. Therefore, commercially available scanners meet this FBI standard. Conversely, high resolution scanners are not commercially available. Commercial use of high resolution scanners, especially in arenas which contemplate use by numerous individuals such as commercial transactions at retail points of sale, remains economically unfeasible.
It will be appreciated from the foregoing that there is a need for a method and apparatus that can reduce the false accept rate and false reject rate of a fingerprint identification system using commercially available biometric scanners.
Yet another need is to identify an individual from an examination of their pore and macrofeature information using commercially available fingerprint scanners.
Yet another need is to reduce the number of false accepts resulting from random pore matching, especially for individuals with high pore density.
Yet another need is for a fingerprint identification system which eliminates fraud attributed to impostor fingerprints.
SUMMARY
In view of the many limitations and drawbacks of previous fingerprint identification systems, and in view of the many long-felt needs associated therewith, it is a first general object of the invention to provide a method and apparatus that can reduce the false accept rate and false reject rate of a fingerprint identification system using commercially available biometric scanners.
It is a further object of the present invention to provide a mechanism to identify an individual from an examination of their pore and macrofeature information using commercially available fingerprint scanners.
Yet another object of the invention is to provide a mechanism to reduce the number of false accepts resulting from random pore matching, especially those individuals with high pore density.
Yet another object is to provide a fingerprint identification system which eliminates fraud attributed to impostor fingerprints.
The invention meets these objects by providing a method of identifying an individual from examination of their natural physical characteristics. Th
Lapsley Philip Dean
Lee Jonathan Alexander
Pare, Jr. David Ferrin
Indivos Corporation
Kamarei Ali
Mehta Bhavesh
LandOfFree
Association of finger pores and macrofeatures for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Association of finger pores and macrofeatures for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Association of finger pores and macrofeatures for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2913838