Integrated segmentation and recognition of handwritten numerals: comparison of classification algorithms
Tóm tắt
In integrated segmentation and recognition (ISR) of handwritten character strings, the underlying classifier is desired to be accurate in character classification and resistant to non-character patterns (also called garbage or outliers). This paper compares the performance of a number of statistical and neural classifiers in ISR. Each classifier has some variations depending on learning method: maximum likelihood estimation (MLE), discriminative learning (DL) under the minimum square error (MSE) or minimum classification error (MCE) criterion, or enhanced DL (EDL) with outlier samples. A heuristic pre-segmentation method is proposed to generate candidate cuts and character patterns. The methods were tested on the 5-digit Zip code images in CEDAR CDROM-1. The results show that training with outliers is crucial for neural classifiers in ISR. The best result was given by the learning quadratic discriminant function (LQDF) classifier.
Từ khóa
#Handwriting recognition #Chromium #Conferences #Image segmentationTài liệu tham khảo
10.2307/2289860
10.1016/S0031-3203(97)00050-2
rainer, 1972, An iterative procedure for the polygonal appro ximationof plane closed curv es, Computer Gr Aphics and Image Pvcessing, 1, 244, 10.1016/S0146-664X(72)80017-0
10.1109/78.175747
liu, 2002, Learning quadratic discriminant function for handwritten character recognition, Proc 16th ICPR
uly ackv, 2001, Probabilistic model for segmentation based w ord recognition with lexicon, Proc 6th ICDAR, 164
10.1109/TPAMI.1987.4767881
10.1016/0031-3203(91)90094-L
liu, 2002, Fujisaw a,P erformance equation of pattern classifiers for handwritten character recognition, Int, J Do CumentA Nalysis and R Ecogni-tion, 4, 191
lee, 1999, Integrated segmentation and recognition of handwritten numerals with cascade neural net w ork jeee trans, System Man and Cybernetics, 29, 285, 10.1109/5326.760572
10.1109/21.260675
10.1016/S0031-3203(00)00018-2
10.1109/5.156471
10.1109/ICPR.1988.28462
yaeger, 1997, Effective training of a neural net w orkcharacter classifier for w ord recognition, A Dvanc Es in Neud Information Processing Systems 9
10.1162/neco.1993.5.3.367
krefiel, 1997, Pattern Classification Techniques Based on Function Approximation, 49
wettschereck, 1992, Dietterich, Improving the performance of radial basis function networks by learning center locations, A Dvanes in Neural Information Processing Systems 4, 11331140