Image analysis – Pattern recognition – Template matching
Reexamination Certificate
1999-05-17
2003-07-15
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Pattern recognition
Template matching
C382S119000, C382S160000, C382S190000, C382S215000, C382S228000, C704S239000, C704S240000, C704S241000, C704S256000
Reexamination Certificate
active
06594392
ABSTRACT:
BACKGROUND
1. Field of the Invention
This invention relates to computer architecture. In particular, the invention relates to processor-based pattern recognition.
2. Description of Related Art
Recognizing an unknown pattern such as speech or handwriting can be accomplished by determining how closely the features of the unknown pattern match to the features of a known, ideal pattern. The features of both the known and unknown patterns can be represented by a multi-dimensional vector in a feature space, with one feature for each dimension.
The comparison, or matching, of the feature vectors between the unknown pattern and the known pattern(s) is inherently a time-consuming process. When the dimensionality of the feature space becomes large, the computations become prohibitively high. In many applications where real-time processing is desired, it is preferable to perform these computations as fast as possible.
In addition, the comparison or matching process typically follow rigid and fixed calculations. Contextual information in forms of weighting factors or probability density function are not fully exploited. The result is that the classification is not robust and flexible, resulting in high classification errors.
Therefore there is a need in the technology to provide a robust and flexible method to perform pattern recognition at a high processing rate.
SUMMARY
The present invention is a method and apparatus to determine a similarity measure between first and second patterns. First and second storages store first and second feature vectors which represent the first and second patterns, respectively. A similarity estimator is coupled to the first and second storages to compute a similarity probability of the first and second feature vectors using a piecewise linear probability density function (PDF). The similarity probability corresponds to the similarity measure.
REFERENCES:
patent: 4608708 (1986-08-01), Watanabe
patent: 4727504 (1988-02-01), Van Broekhoven
patent: 5097509 (1992-03-01), Lennig
patent: 5129002 (1992-07-01), Tsuboka
patent: 5459668 (1995-10-01), Dogan et al.
patent: 5459798 (1995-10-01), Bailey et al.
patent: 5479576 (1995-12-01), Watanabe et al.
patent: 5710866 (1998-01-01), Alleva et al.
patent: 5737488 (1998-04-01), Iso
patent: 5822729 (1998-10-01), Glass
patent: 5857169 (1999-01-01), Seide
patent: 5912989 (1999-06-01), Watanabe
patent: 5991442 (1999-11-01), Yamada et al.
patent: 6067369 (2000-05-01), Kamei
patent: 6078884 (2000-06-01), Downey
patent: 6137909 (2000-10-01), Greineder et al.
patent: 6226409 (2001-05-01), Cham et al.
patent: 6449612 (2002-09-01), Bradley et al.
Desire Gregory
Mehta Bhavesh M.
LandOfFree
Pattern recognition based on piecewise linear probability... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Pattern recognition based on piecewise linear probability..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Pattern recognition based on piecewise linear probability... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3059654