Efficient implementation of fingerprint verification for mobile embedded systems using fixed-point arithmetic

  • Authors:

  • T. Y. Tang;Y. S. Moon;K. C. Chan

  • Affiliations:

  • The Chinese University of Hong Kong, Shatin, N.T. Hong Kong;The Chinese University of Hong Kong, Shatin, N.T. Hong Kong;The Chinese University of Hong Kong, Shatin, N.T. Hong Kong

  • Venue:
  • Proceedings of the 2004 ACM symposium on Applied computing
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Fingerprint sensors are getting small enough to be included in mobile devices to enable fingerprint verification be employed as an authentication tool when using the mobile devices for secure transactions. Fingerprint verification, however, is a computing intensive technology that requires a lot of floating-point computation. Unfortunately, the embedded processors in most mobile devices do not support floating-point hardware. In this paper, we present the implementation of a fingerprint verification process in an embedded system environment based the StrongArm processor and the embedded Linux operating system. The success of the implementation relies on the use of a fixed-point arithmetic only. The fingerprint verification component, the fixed-point component as well as the technique employed to pair up the two components are described in details. In particular, we estimated the required precisions in the fixed-point representations before conducting experiments. Through our results, we further show that not only the fixed-point implementation achieves the goal of significant speed improvement but is almost as reliable as the floating-point counterparts.