抽象-（ abstract-）.docVIP

Hybrid Fingerprint Recognition Junichi Sakamoto DDS http://www.dds.co.jp/en/index.html sakamoto_junichi@dds.co.jp Stephen Miller Aim8 Product Development smiller@ Draft 7 July 2008 Abstract: Minutia based fingerprint recognition algorithms have been widely accepted as a standard for single finger recognition applications. This technology has proved to be a reliable form of enrollment and matching in a corporate environment under idea circumstances. However, the difficulties associated with race, age, application ergonomics, and user interaction often lead to difficult enrollments with verifications requiring multiple attempts. Japan-based biometric company DDS has addressed many of these issues by developing an algorithm, called Spectrum Analysis, which evaluates the fingerprint image through horizontal slices. This proprietary technology has proven easier for enrollment and verification with difficult images. However, the technology does require strict adherence to the ergonomic requirements. To produce the most accurate results under good and difficult environments the company combined two technologies together. This Hybrid Fingerprint technology utilizes a technique called “Shading” where information or lack of information from both algorithms is used to form a decision. This paper addresses the technical aspects and benefits of a hybrid fingerprint recognition algorithm. Introduction Minutiae based fingerprint algorithms require highly accurate templates for enrollment. Many fingers are difficult to enroll due to issues in race, sex, age, and occupation [1]. Moreover, obtaining quality templates in a non-perfect world of poor image quality and application ergonomics can require multiple attempts by trained and untrained users [2]. In a consumer based model, this can lead to increased support costs, costs of sales, and user resistance to technology acceptance. In attempt to address many of the issues there have been numerous recognition technol