Design and acceleration of phase-only filter-based optical pattern recognition for fingerprint identification

We present the use of phase-only filter-based correlation for fingerprint pattern identification. The main advantage of this approach is that it is distortion tolerant and can be realized in optical or electronic parallel hardware. Given that real-world fingerprints are almost never perfect, distort...

Full description

Saved in:
Bibliographic Details
Published in:Optical Engineering 2009-11, Vol.48 (11), p.117206-1172011
Main Authors: Rice, Kenneth L, Taha, Tarek M, Chowdhury, Arshad M, Awwal, Abdul A. S, Woodard, Damon L
Format: Article
Language:English
Subjects:
Acceleration
Algorithms
Correlation
Distortion
field programmable gate array (FPGA) acceleration
Field programmable gate arrays
fingerprint identification
Fingerprints
Hardware
image correlation
phase-only filter
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present the use of phase-only filter-based correlation for fingerprint pattern identification. The main advantage of this approach is that it is distortion tolerant and can be realized in optical or electronic parallel hardware. Given that real-world fingerprints are almost never perfect, distortion tolerance can prove to be very important for this application. Our results indicate that the algorithm can identify prints with 58 of the data missing on average. With large fingerprint databases, identification can be a computationally challenging task. The high parallelism in the phase-only correlation filter makes it ideally suited to field programmable gate array (FPGA)-based hardware acceleration. We examine the FPGA-based acceleration of the fingerprint algorithm. On a Xilinx Virtex II Pro FPGA, we achieve speedups of about 47 times over an optimized C implementation of the algorithm on a AMD Opteron processor. Our FPGA implementation is optimized to allow efficient processing of large databases.
ISSN:0091-3286
1560-2303
DOI:10.1117/1.3265549