Share Email Print

Proceedings Paper

Use of discrete-state Markov process for Chinese-character recognition
Author(s): Bor-Shenn Jeng; Chun-Hsi Shih; San-Wei Sun; Tieh-Min Wu; Bing-Shan Chien; Ming-Wen Chang
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

In this paper, an intelligent optical Chinese character recognition system using discrete-state Markov process has been developed to solve the input problem of Chinese characters. The doubly stochastic process encodes the distortion and similarity among patterns of a class through a stochastical and evaluating approach. A simple feature extraction is employed in our experiments, which is the histogram of the projected profiles. A character class is modeled by a 7-state Markov process, while the recognition rate for 5401 multi-font character classes reaches 97% on average for inside test and 92% on average for outside test.

Paper Details

Date Published: 1 September 1990
PDF: 8 pages
Proc. SPIE 1360, Visual Communications and Image Processing '90: Fifth in a Series, (1 September 1990); doi: 10.1117/12.24181
Show Author Affiliations
Bor-Shenn Jeng, National Central Univ. and Ministry of Communications (Taiwan)
Chun-Hsi Shih, Ministry of Communications (Taiwan)
San-Wei Sun, Ministry of Communications (Taiwan)
Tieh-Min Wu, Ministry of Communications (Taiwan)
Bing-Shan Chien, Ministry of Communications (Taiwan)
Ming-Wen Chang, National Central Univ. (Taiwan)

Published in SPIE Proceedings Vol. 1360:
Visual Communications and Image Processing '90: Fifth in a Series
Murat Kunt, Editor(s)

© SPIE. Terms of Use
Back to Top