Image analysis – Learning systems – Trainable classifiers or pattern recognizers
Patent
1992-08-18
1996-09-10
Couso, Jose L.
Image analysis
Learning systems
Trainable classifiers or pattern recognizers
382197, 382224, G06K 962, G06K 948
Patent
active
055553176
ABSTRACT:
A system that classifies or identifies hand printed or machine printed alphanumeric characters using an iterated least squares polynomial discriminant method is disclosed. Super-sampling of labeled training characters is performed prior to feature extraction. Feature extraction subsamples individual character images. The feature vectors and associated labelings are used to create a weight matrix. During iteration, the weight matrix, to be subsequently used for identification, is modified by determining which characters are incorrectly classified, or classified with too small a confidence, and replicating those characters during training to strengthen the correct classification. The correct classification is also strengthened by using negative feedback to inhibit incorrect classification by using negative values in the target vectors. The speed of the learning process is enhanced by maintaining intermediate matrices and step wise increasing the amount of each feature vector and the size of the sample set used during training. The memory efficiency utilization is enhanced by maintaining and modifying intermediate matrices and compressing the sparse binary features vectors.
REFERENCES:
patent: 5003490 (1991-03-01), Castelaz et al.
patent: 5019903 (1991-05-01), Dougall et al.
patent: 5058180 (1991-10-01), Khan
patent: 5060278 (1991-10-01), Fukumizu
patent: 5060279 (1991-10-01), Crawford et al.
patent: 5063601 (1991-11-01), Hayduk
patent: 5075871 (1991-12-01), Weidman
patent: 5159644 (1992-10-01), Martin et al.
patent: 5161204 (1992-11-01), Hutcheson et al.
P. Gallinari, S. Thiria, F. Badran, and F. Fogelman-Soulie, "On the Relations Between Discriminant Analysis and Multilayer Perceptions", Neural Networks, vol. 4, pp. 349-360, 1991.
William H. Press, Brian P. Flannery, Saul A. Teukolsky, William T. Vetterling, "Numerical Recipes in C.", Cambridge University Press, 1988.
G. W. Stewart, "Introduction to Matrix Computation", Academic Press, 1973, pp. 217-230.
Uma Srinivasan, "Polynomial Discriminant Method For Handwritten Digit Recognition", SUNY Buffalo Technical Report, Dec. 14, 1989.
Andrew R. Webb and David Lowe, "The Optimised Internal Representation of Multiplayer Classifier Networks Performs Nonlinear Discriminant Analysis", Neural Networks, vol. 3, pp. 367-375, 1990.
Charles C. Tappert, Ching Y. Suen, and Toru Wakahara, "The State of the Art in On-Line Handwriting Recoginition", IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 12, No. 8, pp. 787-808, Aug. 1990.
Bella Matthew C.
Couso Jose L.
Eastman Kodak Company
Owens Raymond L.
LandOfFree
Supervised training augmented polynomial method and apparatus fo does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Supervised training augmented polynomial method and apparatus fo, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Supervised training augmented polynomial method and apparatus fo will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1327001