Synthesizing handwritten characters using naturalness learning

Ján Dolinský; Hideyuki Takagi

doi:10.1109/ICCCYB.2007.4402023

Synthesizing handwritten characters using naturalness learning

Ján Dolinský, Hideyuki Takagi

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

12 Citations (Scopus)

Abstract

In this paper we show how to synthesize handwritten characters using a proposed system for naturalness learning. We begin by explaining what we mean by naturalness and then show that in many characters, certain properties of font character strokes does not have a linear relation with this naturalness. This observation inspires the idea of using nonlinear techniques to model the naturalness in order to generate handwriting of a unique, personalized, form. Several techniques for achieving this were tested. Surprisingly, RNN with a recurrent output layer performed the best at generating characters very similar to a person's handwriting.

Original language	English
Title of host publication	ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings
Pages	101-106
Number of pages	6
DOIs	https://doi.org/10.1109/ICCCYB.2007.4402023
Publication status	Published - Dec 1 2007
Event	ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics - Gammarth, Tunisia Duration: Oct 19 2007 → Oct 21 2007

Publication series

Name	ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings

Other

Other	ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics
Country/Territory	Tunisia
City	Gammarth
Period	10/19/07 → 10/21/07

All Science Journal Classification (ASJC) codes

Computer Networks and Communications
Control and Systems Engineering

Access to Document

10.1109/ICCCYB.2007.4402023

Cite this

Synthesizing handwritten characters using naturalness learning. / Dolinský, Ján; Takagi, Hideyuki.
ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings. 2007. p. 101-106 4402023 (ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Dolinský, J & Takagi, H 2007, Synthesizing handwritten characters using naturalness learning. in ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings., 4402023, ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings, pp. 101-106, ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Gammarth, Tunisia, 10/19/07. https://doi.org/10.1109/ICCCYB.2007.4402023

@inproceedings{bebcaf3db00b4aad959654da9d63cc99,

title = "Synthesizing handwritten characters using naturalness learning",

abstract = "In this paper we show how to synthesize handwritten characters using a proposed system for naturalness learning. We begin by explaining what we mean by naturalness and then show that in many characters, certain properties of font character strokes does not have a linear relation with this naturalness. This observation inspires the idea of using nonlinear techniques to model the naturalness in order to generate handwriting of a unique, personalized, form. Several techniques for achieving this were tested. Surprisingly, RNN with a recurrent output layer performed the best at generating characters very similar to a person's handwriting.",

author = "J{\'a}n Dolinsk{\'y} and Hideyuki Takagi",

year = "2007",

month = dec,

day = "1",

doi = "10.1109/ICCCYB.2007.4402023",

language = "English",

isbn = "1424411467",

series = "ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings",

pages = "101--106",

booktitle = "ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings",

note = "ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics ; Conference date: 19-10-2007 Through 21-10-2007",

}

TY - GEN

T1 - Synthesizing handwritten characters using naturalness learning

AU - Dolinský, Ján

AU - Takagi, Hideyuki

PY - 2007/12/1

Y1 - 2007/12/1

N2 - In this paper we show how to synthesize handwritten characters using a proposed system for naturalness learning. We begin by explaining what we mean by naturalness and then show that in many characters, certain properties of font character strokes does not have a linear relation with this naturalness. This observation inspires the idea of using nonlinear techniques to model the naturalness in order to generate handwriting of a unique, personalized, form. Several techniques for achieving this were tested. Surprisingly, RNN with a recurrent output layer performed the best at generating characters very similar to a person's handwriting.

AB - In this paper we show how to synthesize handwritten characters using a proposed system for naturalness learning. We begin by explaining what we mean by naturalness and then show that in many characters, certain properties of font character strokes does not have a linear relation with this naturalness. This observation inspires the idea of using nonlinear techniques to model the naturalness in order to generate handwriting of a unique, personalized, form. Several techniques for achieving this were tested. Surprisingly, RNN with a recurrent output layer performed the best at generating characters very similar to a person's handwriting.

UR - http://www.scopus.com/inward/record.url?scp=47349132104&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47349132104&partnerID=8YFLogxK

U2 - 10.1109/ICCCYB.2007.4402023

DO - 10.1109/ICCCYB.2007.4402023

M3 - Conference contribution

AN - SCOPUS:47349132104

SN - 1424411467

SN - 9781424411467

T3 - ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings

SP - 101

EP - 106

BT - ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics, Proceedings

T2 - ICCC 2007 - 5th IEEE International Conference on Computational Cybernetics

Y2 - 19 October 2007 through 21 October 2007

ER -

Synthesizing handwritten characters using naturalness learning

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this