LEREC - A NN/HMM HYBRID FOR ONLINE HANDWRITING RECOGNITION

被引：66

作者：

BENGIO, Y ^{[1
]}

LECUN, Y ^{[1
]}

NOHL, C ^{[1
]}

BURGES, C ^{[1
]}

机构：

[1] UNIV MONTREAL,DEPT IRO,MONTREAL,PQ H3C 3J7,CANADA

来源：

NEURAL COMPUTATION | 1995年 / 7卷 / 06期

关键词：

D O I：

10.1162/neco.1995.7.6.1289

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We introduce a new approach for on-line recognition of handwritten words written in unconstrained mixed style. The preprocessor performs a word-level normalization by fitting a model of the word structure using the EM algorithm. Words are then coded into low resolution ''annotated images'' where each pixel contains information about trajectory direction and curvature. The recognizer is a convolution network that can be spatially replicated. From the network output, a hidden Markov model produces word scores. The entire system is globally trained to minimize word-level errors.

引用

页码：1289 / 1303

页数：15

共 13 条

[1] GLOBAL OPTIMIZATION OF A NEURAL NETWORK-HIDDEN MARKOV MODEL HYBRID
BENGIO, Y
DEMORI, R
FLAMMIA, G
KOMPE, R
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 1992, 3 (02): : 252 - 259
[2] BURGES C, 1992, P INT JOINT C NEUR N, V3, P165
[3] DESIGN OF A NEURAL NETWORK CHARACTER RECOGNIZER FOR A TOUCH TERMINAL
GUYON, I
ALBRECHT, P
LECUN, Y
DENKER, J
HUBBARD, W
[J]. PATTERN RECOGNITION, 1991, 24 (02) : 105 - 119
[4] KEELER J, 1991, ADV NEURAL INFORMATI, V3, P557
[5] Le Cun Y., 1989, CRGTR894
[6] Backpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y.
Boser, B.
Denker, J. S.
Henderson, D.
Howard, R. E.
Hubbard, W.
Jackel, L. D.
[J]. NEURAL COMPUTATION, 1989, 1 (04) : 541 - 551
[7] LeCun Y., 1989, P ADV NEURAL INFORM, P396
[8] LeCun Y, 1986, DISORDERED SYSTEMS B, P233, DOI DOI 10.1007/978-3-642-82657-3_24
[9] MATAN O, 1992, ADV NEUR IN, V4, P488
[10] ON A MODEL-ROBUST TRAINING METHOD FOR SPEECH RECOGNITION
NADAS, A
NAHAMOO, D
PICHENY, MA
[J]. IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1988, 36 (09): : 1432 - 1436

← 1 2 →