Image analysis – Image transformation or preprocessing – Correlation
Reexamination Certificate
1999-04-27
2003-05-20
Mehta, Bhavesh (Department: 2625)
Image analysis
Image transformation or preprocessing
Correlation
C382S280000, C382S284000
Reexamination Certificate
active
06567566
ABSTRACT:
BACKGROUND OF THE INVENTION
The binary joint transform correlator (BJTC) can easily be applied to fingerprint identification or recognition; see K. H. Fielding, J. L. Horner, and C. K. Makekau, “Optical fingerprint identification by binary joint transform correlation,”
Opt. Eng.
30, 1958-1961, (1991); B. Javidi, and J. L. Horner., “Optical pattern recognition for validation and security verification,” Opt. Eng. 33, 1752-1756, (1994).
Since fingerprint recognition involves comparison of a pair of complicated images, it is an ideal application for optical correlators. While it is not clear that optical processing will ever yield a sizable speed increase over digital parallel processing, using a small digital computer with an optical coprocessor clearly has the potential for reduced size, weight, and power consumption over a digital multiprocessor computer with comparable performance for this application. To assess performance of a recognition system, the ability to discriminate matched prints from unmatched prints should be studied from a statistical viewpoint: see T. J. Grycewicz, “Fingerprint recognition using the binary nonlinear joint transform correlator,”
Optoelectronic Devices and Systems for Processing,
Bahram Javidi and Kristina M. Johnson, ed.,
Critical Reviews of Optical Science and Technology
, Vol. CR65, SPIE Press c. 1996.
Matched prints, referred to herein, are images of the same finger taken at different times, and unmatched prints are images of different fingers. To assure a low probability of false alarm and a low probability of false pass, it is necessary that the largest output peaks for unmatched prints be smaller than the smallest peaks seen for matched prints, in the presence of reasonable distortions from rotation, cuts, abrasions, stretching of the skin, or dirt. The performance metric used here was the probability of passing a false print under the constraint of a constant false alarm rate (CFAR).
It would be desirable to improve the performance of a single spatial light modulator (SLM) BJTC fingerprint correlator; see F. T. S. Yu, et al., “Adaptive real-time pattern recognition using a liquid crystal TV based joint transform correlator”,
Appl. Opt.
26, 1370 (1984).
A number of techniques have been presented for improving binary joint transform correlator (BJTC) performance, and have been shown to improve performance of fingerprint recognition systems. Frame subtraction is one; see T. J. Grycewicz and B. Javidi, “Experimental comparison of binary joint transform correlators used for fingerprint identification,”
Opt. Eng.
35, 2519-2525 (1996); and processing of a partial Fourier plane is another; see T. J. Grycewicz, “Fingerprint recognition using the binary nonlinear joint transform correlator,”
Optoelectronic Devices and Systems for Processing,
Bahram Javidi and Kristina M. Johnson, ed.,
Critical Reviews of Optical Science and Technology
, Vol. CR65. SPIE Press c. 1996.
These additional techniques can be utilized to further improve the performance of the BJTC fingerprint recognition system.
BRIEF SUMMARY OF A PREFERRED EMBODIMENT OF THE INVENTION
Noise in the Fourier plane of a BJTC is screened out by applying a binarization threshold which alternates between even and odd rows in the Fourier plane. Use of such alternating threshold reduces the effect of low amplitude Fourier plane signals on the output, and improves the system capability to discriminate between weak correlations and noise by a factor of greater than three. To counter output amplitude drift in the BJTC system, the output peak heights are normalized, to improve performance by a factor of almost three. When the two techniques are used together, a combined dramatic performance improvement to be described is attained.
REFERENCES:
patent: 5477382 (1995-12-01), Pernick
patent: 5485312 (1996-01-01), Horner et al.
patent: 5629802 (1997-05-01), Clark
patent: 5815597 (1998-09-01), Horner et al.
patent: 6373661 (2002-04-01), Tokuyama et al.
Mehta Bhavesh
Patel Kanji
The United States of America as represented by the Secretary of
USAF ESC/JA
LandOfFree
Techniques to improve binary joint transform correlator,... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Techniques to improve binary joint transform correlator,..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Techniques to improve binary joint transform correlator,... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3004456