Accelerating Large Character Set Recognition using Pivots

Authors:
Yiping Yang;Ondrej Velek;Masaki Nakagawa
Affiliations:
-;-;-
Venue:
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
Year:
2003

Citing 5
Cited 3

Modified Quadratic Discriminant Functions and the Application to Chinese Character Recognition

IEEE Transactions on Pattern Analysis and Machine Intelligence
On speeding candidate selection in handprinted Chinese character recognition

Pattern Recognition
Precise Candidate Selection for Large Character Set Recognition by Confidence Evaluation

IEEE Transactions on Pattern Analysis and Machine Intelligence
On-line Handwritten Character Pattern Database Sampled in a Sequence of Sentences without any Writing Instructions

ICDAR '97 Proceedings of the 4th International Conference on Document Analysis and Recognition
Collection and Analysis of On-line Handwritten Japanese Character Patterns

ICDAR '01 Proceedings of the Sixth International Conference on Document Analysis and Recognition

Camera based mixed-lingual card reader for mobile device

ICDAR '05 Proceedings of the Eighth International Conference on Document Analysis and Recognition
An efficient candidate set size reduction method for coarse-classification in Chinese handwriting recognition

SACH'06 Proceedings of the 2006 conference on Arabic and Chinese handwriting recognition
A robust method for coarse classifier construction from a large number of basic recognizers for on-line handwritten Chinese/Japanese character recognition

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a method to accelerate characterrecognition of a large character set by employing pivotsinto the search space. We divide the feature space ofcharacter categories into smaller clusters and derive thecentroid of each cluster as a pivot. Given an input pattern,it is compared with all the pivots and only a limited numberof clusters whose pivots have higher similarities (orsmaller distances) to the input pattern are searched forwith the result that we can accelerate the recognition speed.This is based on the assumption that the search space is adistance space. The method has been applied topre-classification of a practical off-line Japanesecharacter recognizer with the result that thepre-classification time is reduced to 61 % while keeping itspre-classification recognition rate up to 40 candidates asthe same as the original 99.6% and the total recognitiontime is reduced to 70% of the original time withoutsacrificing the recognition rate at all. If we sacrifice thepre-classification rate from 99.6% to 97.7%, then its timeis reduced to 35% and the total recognition time is reducedto 51.5% with recognition rate as 96.3% from 98.3%.