Advances In Handwriting RecognitionSeong-whan Lee World Scientific, 1999. gada 1. jūn. - 600 lappuses Advances in Handwriting Recognition contains selected key papers from the 6th International Workshop on Frontiers in Handwriting Recognition (IWFHR '98), held in Taejon, Korea from 12 to 14, August 1998. Most of the papers have been expanded or extensively revised to include helpful discussions, suggestions or comments made during the workshop. |
No grāmatas satura
1.–5. rezultāts no 11.
29. lappuse
... bigrams. We also find that preserving a theme in the training text causes relatively little harm to coverage or recognition accuracy. 1 Introduction Accurate, automatic, writer-independent recognition of unconstrained, largevocabulary ...
... bigrams. We also find that preserving a theme in the training text causes relatively little harm to coverage or recognition accuracy. 1 Introduction Accurate, automatic, writer-independent recognition of unconstrained, largevocabulary ...
30. lappuse
... bigram compactness, includes at least one occurrence of the character pairs most-commonly occurring in general use. 2 Recognition System. Overview Experiments are performed using the IBM handwriting recognizer in the IBM. Ink. Manager ...
... bigram compactness, includes at least one occurrence of the character pairs most-commonly occurring in general use. 2 Recognition System. Overview Experiments are performed using the IBM handwriting recognizer in the IBM. Ink. Manager ...
31. lappuse
Seong-whan Lee. represents at least 0.03% of bigrams in general usage”. Thus, unigram balance is prioritized toward repetition without regard to context or frequency of occurrence, while bigram compactness does the opposite, at the ...
Seong-whan Lee. represents at least 0.03% of bigrams in general usage”. Thus, unigram balance is prioritized toward repetition without regard to context or frequency of occurrence, while bigram compactness does the opposite, at the ...
32. lappuse
... bigrams, we use a universe corpus drawn primarily from newspapers and office correspondence. This corpus consists of 338 million words of text, with over 206,000 distinct words. Character bigram statistics were taken, treating word ...
... bigrams, we use a universe corpus drawn primarily from newspapers and office correspondence. This corpus consists of 338 million words of text, with over 206,000 distinct words. Character bigram statistics were taken, treating word ...
33. lappuse
... bigram compactness largely to get the bigrams “ys” and “00”. In each case, the trivia fact remains essentially intact, and the wording becomes only slightly awkward. Much more text manipulation can be done when content need not be ...
... bigram compactness largely to get the bigrams “ys” and “00”. In each case, the trivia fact remains essentially intact, and the wording becomes only slightly awkward. Much more text manipulation can be done when content need not be ...
Saturs
17 | |
HANDWRITTEN FORM PROCESSING | 79 |
HANDWRITTEN WORD RECOGNITION | 151 |
SEGMENTATION | 223 |
ORIENTAL SCRIPT PROCESSING | 275 |
NUMERAL RECOGNITION | 357 |
EMERGING TECHNIQUES | 437 |
APPLICATIONS | 517 |
Citi izdevumi - Skatīt visu
Bieži izmantoti vārdi un frāzes
algorithm applied approach B-spline bankcheck bigram character candidates character line character recognition character segmentation classifier combination computed connected components contour corresponding courtesy amount database described detected diacriticals dictionary digit distance error rate evaluation example experimental results experiments feature extraction feature set feature vectors Figure function fuzzy graph grapheme handwriting recognition handwritten character handwritten numerals handwritten words Hidden Markov Models horizontal Hough transform hypotheses integrated Kanji learning legal amount letters lexicon line segments matching matrix nat-ja neural network node obtained off-line on-line handwriting recognition optimal output paper parameters Pattern Recognition performance pixels points preprocessing probability problem Proc proposed method prototype quantization recognition rate recognition results recognition system recognizer samples sequence shown speech recognition step string stroke structure Suen Table technique template threshold touching type vector quantizer wavelet word recognition writing