Presentation | 2018-12-13 情景内文字のCNNによる拡大 Toshiki Nakamura, Seiichi Uchida, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | The purpose of this research is magnifying scene texts using convolutional neural network (CNN) with end-to-end. We trained Encoder-Decoder type CNN composed of convolution/deconvolution using images magnified scene text without changing the center of gravity as teacher image. Also, we tried a method of magnifying scene texts by dividing the task into 4 of text erasing, text extraction, text magnifying, and image merging, and learning 4 CNNs for each, and combining them. Furthermore, we merging 1 CNN these 4 CNNs and fine tuned by reusing weights. In addition, we compare these results with text magnifying method using character detection and evaluated it quantitatively. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | |
Paper # | PRMU2018-76 |
Date of Issue | 2018-12-06 (PRMU) |
Conference Information | |
Committee | PRMU |
---|---|
Conference Date | 2018/12/13(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Shinichi Sato(NII) |
Vice Chair | Yoshihisa Ijiri(Omron) / Toru Tamaki(Hiroshima Univ.) |
Secretary | Yoshihisa Ijiri(NEC) / Toru Tamaki(Osaka Univ.) |
Assistant | Go Irie(NTT) / Yoshitaka Ushiku(OSX) |
Paper Information | |
Registration To | Technical Committee on Pattern Recognition and Media Understanding |
---|---|
Language | JPN-ONLY |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | |
Sub Title (in English) | |
Keyword(1) | |
Keyword(2) | |
Keyword(3) | |
1st Author's Name | Toshiki Nakamura |
1st Author's Affiliation | Kyushu University(Kyushu Univ.) |
2nd Author's Name | Seiichi Uchida |
2nd Author's Affiliation | Kyushu University(Kyushu Univ.) |
Date | 2018-12-13 |
Paper # | PRMU2018-76 |
Volume (vol) | vol.118 |
Number (no) | PRMU-362 |
Page | pp.pp.7-12(PRMU), |
#Pages | 6 |
Date of Issue | 2018-12-06 (PRMU) |