Recognizing characters in scene images

Jun Ohya, Akio Shio, Shigeru Akamatsu

Research output: Contribution to journalArticle

168 Citations (Scopus)

Abstract

An effective algorithm for character recognition in scene images is studied. Scene images are segmented into regions by an image segmentation method based on adaptive thresholding. Character candidate regions are detected by observing gray-level differences between adjacent regions. To ensure extraction of multisegment characters as well as single-segment characters, character pattern candidates are obtained by associating the detected regions according to their positions and gray levels. A character recognition process selects patterns with high similarities by calculating the similarities between character pattern candidates and the standard patterns in a dictionary and then comparing the similarities to the thresholds. A relaxational approach to determine character patterns updates the similarities by evaluating the interactions between categories of patterns, and finally character patterns and their recognition results are obtained. Highly promising experimental results have been obtained using the method on 100 images involving characters of different sizes and formats under uncontrolled lighting.

Original languageEnglish
Pages (from-to)214-220
Number of pages7
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume16
Issue number2
DOIs
Publication statusPublished - 1994 Feb
Externally publishedYes

Fingerprint

Character recognition
Glossaries
Image segmentation
Lighting
Character Recognition
Adaptive Thresholding
Character
Image Segmentation
Adjacent
Update
Similarity
Experimental Results
Interaction

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Recognizing characters in scene images. / Ohya, Jun; Shio, Akio; Akamatsu, Shigeru.

In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 16, No. 2, 02.1994, p. 214-220.

Research output: Contribution to journalArticle

Ohya, Jun ; Shio, Akio ; Akamatsu, Shigeru. / Recognizing characters in scene images. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 1994 ; Vol. 16, No. 2. pp. 214-220.
@article{00a6a6a2b9bf46d7bc05712f3fd29bbc,
title = "Recognizing characters in scene images",
abstract = "An effective algorithm for character recognition in scene images is studied. Scene images are segmented into regions by an image segmentation method based on adaptive thresholding. Character candidate regions are detected by observing gray-level differences between adjacent regions. To ensure extraction of multisegment characters as well as single-segment characters, character pattern candidates are obtained by associating the detected regions according to their positions and gray levels. A character recognition process selects patterns with high similarities by calculating the similarities between character pattern candidates and the standard patterns in a dictionary and then comparing the similarities to the thresholds. A relaxational approach to determine character patterns updates the similarities by evaluating the interactions between categories of patterns, and finally character patterns and their recognition results are obtained. Highly promising experimental results have been obtained using the method on 100 images involving characters of different sizes and formats under uncontrolled lighting.",
author = "Jun Ohya and Akio Shio and Shigeru Akamatsu",
year = "1994",
month = "2",
doi = "10.1109/34.273729",
language = "English",
volume = "16",
pages = "214--220",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "2",

}

TY - JOUR

T1 - Recognizing characters in scene images

AU - Ohya, Jun

AU - Shio, Akio

AU - Akamatsu, Shigeru

PY - 1994/2

Y1 - 1994/2

N2 - An effective algorithm for character recognition in scene images is studied. Scene images are segmented into regions by an image segmentation method based on adaptive thresholding. Character candidate regions are detected by observing gray-level differences between adjacent regions. To ensure extraction of multisegment characters as well as single-segment characters, character pattern candidates are obtained by associating the detected regions according to their positions and gray levels. A character recognition process selects patterns with high similarities by calculating the similarities between character pattern candidates and the standard patterns in a dictionary and then comparing the similarities to the thresholds. A relaxational approach to determine character patterns updates the similarities by evaluating the interactions between categories of patterns, and finally character patterns and their recognition results are obtained. Highly promising experimental results have been obtained using the method on 100 images involving characters of different sizes and formats under uncontrolled lighting.

AB - An effective algorithm for character recognition in scene images is studied. Scene images are segmented into regions by an image segmentation method based on adaptive thresholding. Character candidate regions are detected by observing gray-level differences between adjacent regions. To ensure extraction of multisegment characters as well as single-segment characters, character pattern candidates are obtained by associating the detected regions according to their positions and gray levels. A character recognition process selects patterns with high similarities by calculating the similarities between character pattern candidates and the standard patterns in a dictionary and then comparing the similarities to the thresholds. A relaxational approach to determine character patterns updates the similarities by evaluating the interactions between categories of patterns, and finally character patterns and their recognition results are obtained. Highly promising experimental results have been obtained using the method on 100 images involving characters of different sizes and formats under uncontrolled lighting.

UR - http://www.scopus.com/inward/record.url?scp=0028379742&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028379742&partnerID=8YFLogxK

U2 - 10.1109/34.273729

DO - 10.1109/34.273729

M3 - Article

AN - SCOPUS:0028379742

VL - 16

SP - 214

EP - 220

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 2

ER -