Presentation | 2019-10-18 Recovering Dynamic Trajectory from a handwritten character image using a Deep Neural Network Tsubasa Nakamura, Nguyen Tuan Cuong, Masaki Nakagawa, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper presents a method to recover a dynamic trajectory from a handwritten character image using a deep neural network and consider whether the dynamic trajectory is useful for handwritten character recognition. Recognition rate of off-line handwritten characters is lower than that of on-line handwritten characters because off-line character patterns are recognized only by static information from their two-dimensional images. Therefore, the recognition rate of off-line character patterns could be improved if their trajectories were recovered from their images. Moreover, the trajectory recovery may help archaeologists decode degraded character patterns on historical documents. We recover the trajectory by an Encoder-Decoder model with the attention mechanism. In this paper, we evaluate our trajectory recovery qualitatively as well as quantitatively by combining online recognition of the recovered dynamic trajectories with offline recognition of the original character images. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | offline handwritten character pattern / trajectory recevery / deep neural network |
Paper # | PRMU2019-41 |
Date of Issue | 2019-10-11 (PRMU) |
Conference Information | |
Committee | PRMU |
---|---|
Conference Date | 2019/10/18(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Yoichi Sato(Univ. of Tokyo) |
Vice Chair | Toru Tamaki(Hiroshima Univ.) / Akisato Kimura(NTT) |
Secretary | Toru Tamaki(NTT) / Akisato Kimura(OMRON SINICX) |
Assistant | Yusuke Uchida(DeNA) / Takayoshi Yamashita(Chubu Univ.) |
Paper Information | |
Registration To | Technical Committee on Pattern Recognition and Media Understanding |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Recovering Dynamic Trajectory from a handwritten character image using a Deep Neural Network |
Sub Title (in English) | |
Keyword(1) | offline handwritten character pattern |
Keyword(2) | trajectory recevery |
Keyword(3) | deep neural network |
1st Author's Name | Tsubasa Nakamura |
1st Author's Affiliation | Tokyo University of Agriculture and Technology(TUAT) |
2nd Author's Name | Nguyen Tuan Cuong |
2nd Author's Affiliation | Tokyo University of Agriculture and Technology(TUAT) |
3rd Author's Name | Masaki Nakagawa |
3rd Author's Affiliation | Tokyo University of Agriculture and Technology(TUAT) |
Date | 2019-10-18 |
Paper # | PRMU2019-41 |
Volume (vol) | vol.119 |
Number (no) | PRMU-235 |
Page | pp.pp.53-58(PRMU), |
#Pages | 6 |
Date of Issue | 2019-10-11 (PRMU) |