Synthetic Scene Character Generator and Multi-Scale Voting Classifier for Japanese Scene Character Recognition

Fuma Horie, Hideaki Goto

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Scene character recognition is challenging owing to various noise and distortions. In addition, Japanese character recognition requires a large number of training data since thousands of character classes exist in the language. Some researches proposed training data augmentation techniques using synthetic scene character data (SSD)to compensate for the shortage of training data. We proposed multi-scale scheme using multiple dataset consisting of SSD in our previous work to improve recognition accuracy. For further improvement of the scheme, we then proposed Random Filter as SSD generator. This paper enhances the effectiveness of the multi-scale scheme and Random Filter by using larger dataset, JPSC1400, which we have developed by collecting Japanese real scene characters. Experimental results show that the accuracy has been improved from 61.5% to 65.5 % by newly introduced an affine transformation to the SSD generation.

Original languageEnglish
Title of host publication2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018
PublisherIEEE Computer Society
ISBN (Electronic)9781728101255
DOIs
Publication statusPublished - 2019 Feb 4
Event2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018 - Auckland, New Zealand
Duration: 2018 Nov 192018 Nov 21

Publication series

NameInternational Conference Image and Vision Computing New Zealand
Volume2018-November
ISSN (Print)2151-2191
ISSN (Electronic)2151-2205

Conference

Conference2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018
Country/TerritoryNew Zealand
CityAuckland
Period18/11/1918/11/21

Keywords

  • Japanese scene character recognition
  • character recognition
  • ensemble voting classifier
  • scene character dataset
  • synthetic scene character data

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Synthetic Scene Character Generator and Multi-Scale Voting Classifier for Japanese Scene Character Recognition'. Together they form a unique fingerprint.

Cite this