| 日 時 |
平成23年9月30日 （金） 16:00～ 17:30 |
| 会 場 |
九州大学伊都キャンパスウエスト2号館3Ｆ システム情報科学研究院 第1講義室 |
| 演 題 |
『Handwriting Recognition and Innovative Applications』 |
| 講 師 |
German Research Center for Artificial Intelligence (DFKI)
Dr. Marcus Liwicki 氏 |
| 概 要 |
Analysis, recognition, and identification of handwriting is one of the
most important topics of artificial intelligence. Since handwriting is
created by hand motion and acquired by digital pen device or tablet, it
is considered as a temporal motion pattern, like human gesture and speech.
It also has another aspect as an image pattern because it can be shown
as a bitmap pattern on a paper. Consequently, the research on this topic
gives wide fundamental technologies as well as practically useful applications.
In this presentation I will first demonstrate a successful recognition
system for handwritten data. It is based on a new graphical model called
bi-directional and multi-directional long short-term memory networks.
In the second half of my presentation, recent applications in various areas
will be described and demonstrated. Some of these systems have also been
evaluated by an independent group of users who actually use the system. |
| 問合先 |
九州大学大学院システム情報科学研究院情報知能工学部門 内田誠一  |
| 主催・共催 |
【主催】電子情報通信学会 九州支部 |
| その他 |
（聴講料）無料 |
