G - Physics – 06 – K
Patent
G - Physics
06
K
354/133, 354/59
G06K 9/80 (2006.01) G06G 7/48 (2006.01) G06K 9/32 (2006.01) G06K 9/36 (2006.01) G06K 9/44 (2006.01)
Patent
CA 2002542
A highly effective OCR system that combines various techniques to remove meaningless variability and capture meaningful variability on the image before attempting to classify the characters within the image. Specifically, characters to be recognized are scanned to form a pixel image, and the image is scaled and "cleansed." The scaled and cleansed image is deskewed and skeletonized to form a thinned image, and the thinned image is applied to a neural network that extracts features of the image and forms an image map. Features are then combined to form "super features." Thereafter, the image is coarse blocked to reduce the dimensionality of the resulting feature vector. The feature vector is, lastly, applied to a classified that identifies the character.
Baird Henry Spalding
Denker John Stewart
Graf Hans Peter
Henderson Donnie
Howard Richard Edwin
American Telephone And Telegraph Company
Kirby Eades Gale Baker
LandOfFree
Imaged symbol classification does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Imaged symbol classification, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Imaged symbol classification will most certainly appreciate the feedback.
Profile ID: LFCA-PAI-O-1488392