What does scene text tell us?

Seiichi Uchida, Yuto Shinahara

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

Abstract

Scene text is one of the most important information sources for our daily life because it has particular functions such as disambiguation and navigation. In contrast, ordinary document text has no such function. Consequently, it is natural to have a hypothesis that scene text and document text have different characteristics. This paper tries to prove this hypothesis by semantic analysis of texts by word2vec, which is a neural network model to give a vector representation of each word. By the vector representation, we can have the semantic distributions of scene text and document text in Euclidean space and then determine their semantic categories by simple clustering. Experimental study reveals several differences between scene text and document text. For example, it is found that scene text is a semantic subset of document text and several semantic categories are very specific to scene text.

Original languageEnglish
Title of host publication2016 23rd International Conference on Pattern Recognition, ICPR 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4047-4052
Number of pages6
ISBN (Electronic)9781509048472
DOIs
Publication statusPublished - Jan 1 2016
Event23rd International Conference on Pattern Recognition, ICPR 2016 - Cancun, Mexico
Duration: Dec 4 2016Dec 8 2016

Publication series

NameProceedings - International Conference on Pattern Recognition
Volume0
ISSN (Print)1051-4651

Other

Other23rd International Conference on Pattern Recognition, ICPR 2016
CountryMexico
CityCancun
Period12/4/1612/8/16

Fingerprint

Semantics
Navigation
Neural networks

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

Cite this

Uchida, S., & Shinahara, Y. (2016). What does scene text tell us? In 2016 23rd International Conference on Pattern Recognition, ICPR 2016 (pp. 4047-4052). [7900267] (Proceedings - International Conference on Pattern Recognition; Vol. 0). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICPR.2016.7900267

What does scene text tell us? / Uchida, Seiichi; Shinahara, Yuto.

2016 23rd International Conference on Pattern Recognition, ICPR 2016. Institute of Electrical and Electronics Engineers Inc., 2016. p. 4047-4052 7900267 (Proceedings - International Conference on Pattern Recognition; Vol. 0).

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

Uchida, S & Shinahara, Y 2016, What does scene text tell us? in 2016 23rd International Conference on Pattern Recognition, ICPR 2016., 7900267, Proceedings - International Conference on Pattern Recognition, vol. 0, Institute of Electrical and Electronics Engineers Inc., pp. 4047-4052, 23rd International Conference on Pattern Recognition, ICPR 2016, Cancun, Mexico, 12/4/16. https://doi.org/10.1109/ICPR.2016.7900267
Uchida S, Shinahara Y. What does scene text tell us? In 2016 23rd International Conference on Pattern Recognition, ICPR 2016. Institute of Electrical and Electronics Engineers Inc. 2016. p. 4047-4052. 7900267. (Proceedings - International Conference on Pattern Recognition). https://doi.org/10.1109/ICPR.2016.7900267
Uchida, Seiichi ; Shinahara, Yuto. / What does scene text tell us?. 2016 23rd International Conference on Pattern Recognition, ICPR 2016. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 4047-4052 (Proceedings - International Conference on Pattern Recognition).
@inproceedings{1fd265a7673645f4b79adda59d0452f6,
title = "What does scene text tell us?",
abstract = "Scene text is one of the most important information sources for our daily life because it has particular functions such as disambiguation and navigation. In contrast, ordinary document text has no such function. Consequently, it is natural to have a hypothesis that scene text and document text have different characteristics. This paper tries to prove this hypothesis by semantic analysis of texts by word2vec, which is a neural network model to give a vector representation of each word. By the vector representation, we can have the semantic distributions of scene text and document text in Euclidean space and then determine their semantic categories by simple clustering. Experimental study reveals several differences between scene text and document text. For example, it is found that scene text is a semantic subset of document text and several semantic categories are very specific to scene text.",
author = "Seiichi Uchida and Yuto Shinahara",
year = "2016",
month = "1",
day = "1",
doi = "10.1109/ICPR.2016.7900267",
language = "English",
series = "Proceedings - International Conference on Pattern Recognition",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "4047--4052",
booktitle = "2016 23rd International Conference on Pattern Recognition, ICPR 2016",
address = "United States",

}

TY - GEN

T1 - What does scene text tell us?

AU - Uchida, Seiichi

AU - Shinahara, Yuto

PY - 2016/1/1

Y1 - 2016/1/1

N2 - Scene text is one of the most important information sources for our daily life because it has particular functions such as disambiguation and navigation. In contrast, ordinary document text has no such function. Consequently, it is natural to have a hypothesis that scene text and document text have different characteristics. This paper tries to prove this hypothesis by semantic analysis of texts by word2vec, which is a neural network model to give a vector representation of each word. By the vector representation, we can have the semantic distributions of scene text and document text in Euclidean space and then determine their semantic categories by simple clustering. Experimental study reveals several differences between scene text and document text. For example, it is found that scene text is a semantic subset of document text and several semantic categories are very specific to scene text.

AB - Scene text is one of the most important information sources for our daily life because it has particular functions such as disambiguation and navigation. In contrast, ordinary document text has no such function. Consequently, it is natural to have a hypothesis that scene text and document text have different characteristics. This paper tries to prove this hypothesis by semantic analysis of texts by word2vec, which is a neural network model to give a vector representation of each word. By the vector representation, we can have the semantic distributions of scene text and document text in Euclidean space and then determine their semantic categories by simple clustering. Experimental study reveals several differences between scene text and document text. For example, it is found that scene text is a semantic subset of document text and several semantic categories are very specific to scene text.

UR - http://www.scopus.com/inward/record.url?scp=85019087602&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85019087602&partnerID=8YFLogxK

U2 - 10.1109/ICPR.2016.7900267

DO - 10.1109/ICPR.2016.7900267

M3 - Conference contribution

AN - SCOPUS:85019087602

T3 - Proceedings - International Conference on Pattern Recognition

SP - 4047

EP - 4052

BT - 2016 23rd International Conference on Pattern Recognition, ICPR 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -