| Project/Area Number |
26330416
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Learning support system
|
|---|
| Research Institution | Matsue National College of Technology |
|---|
Principal Investigator |
Satoru Kato 松江工業高等専門学校, 情報工学科, 准教授 (40342547)
|
|---|
| Co-Investigator(Renkei-kenkyūsha) |
HORIUCHI Tadashi 松江工業高等専門学校, 電子制御工学科, 教授 (50294129)
|
|---|
| Research Collaborator |
MORITA Hiroaki
ASANO Ryota
|
|---|
| Project Period (FY) |
2014-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
Fiscal Year 2016: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2015: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2014: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
|
|---|
| Keywords | 古文書文字認識 / 自己組織化マップ / マルチテンプレート / 古文書読解支援システム / 並列処理 / PCクラスタ / GPGPU / 認識精度 / サポートベクターマシン / マハラノビス距離 / SOMテンプレート / 認識精度向上 / クラスタリング |
|---|
| Outline of Final Research Achievements |
In this study, several methods of Japanese historical character recognition are examined. It is revealed that the SOM based method called "SOM template" has excellent noise resistance, and multi-template method is good for the balance of trade-off between accuracy and computation time. A prototype of the reading support system for Japanese historical documents is constructed on a tablet terminal. Furthermore, PC-cluster and GPGPU computing method is used for speeding up the SOM learning process and we confirmed that this approach is effective in a large scale of the SOM learning process.
|
