FAST DICTIONARY LOOK-UP FOR CONTEXTUAL WORD RECOGNITION

被引：16

作者：

WELLS, CJ ^{[1
]}

EVETT, LJ ^{[1
]}

WHITBY, PE ^{[1
]}

WHITROW, RJ ^{[1
]}

机构：

[1] TRENT POLYTECH, DEPT COMP, BURTON ST, NOTTINGHAM NG1 4BU, ENGLAND

来源：

PATTERN RECOGNITION | 1990年 / 23卷 / 05期

关键词：

Dictionary; Lexicon; n-grams; Script recognition; Transitional probabilities;

D O I：

10.1016/0031-3203(90)90071-R

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Script recognition systems require the use of context in order to increase the accuracy of their output. One level of context commonly used is that of the surrounding letters. It is demonstrated that the most effective method for utilizing such context is via a dictionary look-up. The present paper describes a data structure for representing a large dictionary (about 60,000 words) which can be searched in real time and uses a practical amount of memory. © 1990.

引用

页码：501 / 508

页数：8

共 28 条

[1] DESIGNING A HANDWRITING READER [J].

BURR, DJ .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1983, 5 (05) :554-559

[2]

CRYSTAL D, 1989, ENCY LANGUAGE

[3] EXPERIMENTS IN CONTEXTUAL RECOGNITION OF CURSIVE SCRIPT [J].

EHRICH, RW ;

KOEHLER, KJ .

IEEE TRANSACTIONS ON COMPUTERS, 1975, C 24 (02) :182-194

[4]

EVETT LJ, 1989, WORDS COMBINE EFFECT

[5]

EVETT LJ, 1988, POSTPROCESSING TECHN

[6]

FARAG RFH, 1979, IEEE T COMPUT, V28, P172, DOI 10.1109/TC.1979.1675310

[7]

FRISCHKOPF LS, 1961, INFORMATION THEORY

[8] CONTEXTUAL WORD RECOGNITION USING PROBABILISTIC RELAXATION LABELING [J].

GOSHTASBY, A ;

EHRICH, RW .

PATTERN RECOGNITION, 1988, 21 (05) :455-462

[9]

HIGGINS CA, 1984, 1ST P INT 84 IFIP C, P140

[10]

Hornby A S., 1988, OXFORD ADV LEARNERS

← 1 2 3 →