Presentation | 2019-03-13 [Poster Presentation] Human Touch Character Font System Yuto Yamamoto, Michiharu Niimi, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | For modern computer society, one may need human touch character font system to enjoy making communication through Internet. An approach to build such font system is to modify the usual font installed in one's computer devices. For this approach, we report hand written HIRAGANA character generation using neural network. As neural network in this paper, a simple auto encoder consisting of input layer, middle layer and output layer, and deep neural network based on GAN (Generative Adversarial Netowrk) called zi2zi are adopted. We use these neural networks as an image modification filter that input data is usual font data and output data is hand written character. We investigate the feasibility of those neural networks to generate hand written characters of HIRAGANA experimentally. For auto encoder, the output image was like rounded character of input rasterized font, and it was not like hand written characters of training data. On the other hand, for zi2zi, we confirmed that the output image is quite like hand written character if we can prepare much training data. However, for untrained data, the output image is almost like hand written character only in some case. In most cases, the output image is roughly equal to a character or was not able to be recognized as a character visually. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Auto Encoder / GAN / hand written character font / machine learning / font generation |
Paper # | EMM2018-106 |
Date of Issue | 2019-03-06 (EMM) |
Conference Information | |
Committee | EMM |
---|---|
Conference Date | 2019/3/13(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | TBD |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Image and Sound Quality, Metrics for Perception and Recognition, Human Auditory and Visual System, etc. |
Chair | Keiichi Iwamura(TUC) |
Vice Chair | Minoru Kuribayashi(Okayama Univ.) / Tetsuya Kojima(NIT,Tokyo College) |
Secretary | Minoru Kuribayashi(NIT, Tokyo) / Tetsuya Kojima(Tyukyo Univ.) |
Assistant | Hiroko Akiyama(NIT, Nagano College) / キタヒロ カネダ(CANON) |
Paper Information | |
Registration To | Technical Committee on Enriched MultiMedia |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Poster Presentation] Human Touch Character Font System |
Sub Title (in English) | Generation of hand written HIRAGANA using Neural Network |
Keyword(1) | Auto Encoder |
Keyword(2) | GAN |
Keyword(3) | hand written character font |
Keyword(4) | machine learning |
Keyword(5) | font generation |
1st Author's Name | Yuto Yamamoto |
1st Author's Affiliation | Kyushu Institute of Technology(KIT) |
2nd Author's Name | Michiharu Niimi |
2nd Author's Affiliation | Kyushu Institute of Technology(KIT) |
Date | 2019-03-13 |
Paper # | EMM2018-106 |
Volume (vol) | vol.118 |
Number (no) | EMM-494 |
Page | pp.pp.77-82(EMM), |
#Pages | 6 |
Date of Issue | 2019-03-06 (EMM) |