JSLコーパス構築を目指して
スポンサーリンク
概要
- 論文の詳細を見る
The recent development of HDD video-recorders and computers has enabled us to record and analyze the visual languages such as Japanese Sign Language (JSL). However, as far as the recorded data of JSL are concerned, they are insufficient in both quality and quantity for our linguistic investigation. Therefore, the purpose of this study is to emphasize the necessity to create a JSL corpus; the aggregation of annotated discourses, which was not translated from the spoken language but signed spontaneously, and metadata which describe the information on the signer. We will show an analysis that the observation of the spontaneously signed representations reveals a part of JSL structure. We will also introduce methodology to build a useful corpus.