Bioindex mechanism for increasing the relative speed of biometric indentification against large population samples

Abstract

A method and system for determining a smaller sample of large data objects to search through for a match to one under test. The data objects are in preferred embodiments biometric templates.

Description

BACKGROUND OF THE INVENTION [0001]1. Field of the Invention[0002]This invention relates to the field of data processing generally, particularly to performing fast searches in data having many large data objects to find matches to an unknown similar large data object. It has particular application to authentication utilizing biometric data and more specifically the method and process of biometric database searching.[0003]2. Background[0004]A significant problem facing biometric identification using a large population is the processing time required to perform a full comparison on the entire population to a presenting person, so as to identify that person. Even when the database processing is optimized and the data is available before the comparison, the biometric comparison operation becomes a timing bottleneck that slows the identification process on large populations to such a degree as to make real time use of such searches unavailable. One way to decrease the time required for bi...

AI Technical Summary

Benefits of technology

[0009]It is an object of this invention to overcome the disadvantages of prior methods by defining an implementation-independent mathematically calculated index which will be used to reliably reduce the population of templates in the overall population of templates that may be required for a full biometric comparison. It is a further object of this invention is to provide a flexible method in which various biometric types, devices, algorithms and storage mechanisms may be used in the solution. It is also a further object of the invention is to present the data in a standard form which may enable different biometric templates to be stored in a common location, i.e. enterprise biometrics storage. It should also be noted that the techniques described herein can be applied to any large data object database and search for matches therein to a particular large data object under consideration.

[0012]In the third step of the preferred embodiment, a Dynamic Key (xd) is generated during the identification process using the same process as the Template Key (xt) generation on either the same biometric machine or another biometric machine. Depending on the biometric collection device, certain portions of the least significant attributes may either be dropped or not generated at all. Also, depending on the biometric collection device, a collection range error is included with the template, increasing the range of keys searched by the amount of the error.

[0013]In the third step of the preferred embodiment, matching of the Template Key (xt) to the Dynamic Key authenticates the user. In the preferred embodiment, the Template Key (xt) database field is reduced by matching the most significant digit of the Dynamic Key to the most significant digit of the Template Keys (xt). The Template Key (xt) search field then contains only those numbers whose most significant digits match the most significant digit of the Dynamic Key. A further narrowing search is performed by matching the next most significant digit of the Dynamic Key with the next most significant digit of the Template Key (xt) in the previously reduced Template Key database field. Subsequent to the search, the Template Key search field contains only those numbers whose most significant and second most significant digits match the most significant and second most significant digits of the Dynamic Key. We could use a range of match, such that for example a 6 could be considered within range of a 4 if the range is exact number plus 2 and minus 1, for example. This would be valuable where there is great variation in a particular significant digit / attribute known for a given biometric. There can be substantial variation from one biometric to another and even within particular biometric reader data template populations and these variables need to be controlled. The invention will work best within a single biometric and single type of reader / template generator combination. It is advised that the key attributes be selected based on known characteristics of the biometric and reader / template generator combination. Note that after the Template keys that fall within range or match the Dynamic Key are found, all those templates in the overall population database can be searched to determine matches to the collected biometric data's template, thus reducing significantly the overall search time needed to locate an exact match.

Problems solved by technology

Also, depending on the biometric collection device, a collection range error is included with the template, increasing the range of keys searched by the amount of the error.

Images