Research for Statistical Analysis of Japanese Sign Language Corpus and its Application to Sign Language Teaching

日语手语语料库统计分析及其在手语教学中的应用研究

基本信息

批准号：
14208017
负责人：
KANDA Kazuyuki
金额：
$ 26.04万
依托单位：
Chukyo University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (A)
财政年份：
2002
资助国家：
日本
起止时间：
2002 至 2005
项目状态：
已结题

项目摘要

The years 2002-2003 : This was the first year and we focused on the former works. We examined the hundred sign utterances gained before and found the errors to correct. We also removed the national errors. We collected the data from The Proficiency Test of Japanese Sign Language as the sign corpus. We found the co-relation between the nodding and other NMS in the end of a clause in the experimental analysis.In the years 2003-2004,through checking the PTJSL data, we found all the expressions are highly controlled and they are clearly separated into words and sentences, which are suitable for our purpose. We created the database of them which is made of the stratum of the phonological, the lexical and the grammatical level. We collected the sign corpus in the dialog situation and analyzed the co-relation among the NMSs.In the years 2004-2005,we added the corpus data and input them into the database. We analyzed the dialog corpus and studied the turn-taking rules. We also had some experiments the controlled signings with the mask to control NMS and with the fists to control the hand shapesIn the years 2005-2006,we analyzed the lexicon in the sign corpus and made the list of the variety of the signs in the same lexical entry, which we expect to set up some standard of JSL lexis. We reported the list of difficult sing to learn for the learners. It will help the sign education as well as the sign recognition system.These findings and results lead to the next step of the research of sign formation and morphology of the sign language.

2002-2003年：这是第一年，我们专注于以前的作品。我们检查了以前获得的一百个迹象说法，并发现了纠正的错误。我们还删除了国家错误。我们从日语手语的能力测试中收集了数据作为标志语料库。我们在实验分析中的子句结束时发现了点头和其他NM之间的共同关系。在2003-2004年，通过检查PTJSL数据，我们发现所有表达式都受到了高度控制，并且它们清楚地分离为单词和句子，适合我们的目的。我们创建了它们的数据库，该数据库由语音，词汇和语法级别的阶层制成。我们在对话情况下收集了标志语料库，并分析了NMS之间的共同关系。在2004-2005年，我们添加了语料库数据并将其输入数据库。我们分析了对话语料库并研究了转弯规则。我们还进行了一些实验，这些实验是用面具的受控签名来控制NMS，并用拳头控制了2005-2006年的手工形状，我们在标志语料库中分析了词典，并将各种标志的列表列出了相同的列表。词汇条目，我们希望它设置一些JSL Lexis标准。我们报告了为学习者学习的难以学习的清单。这将有助于标志教育以及标志识别系统。这些发现和结果导致了符号形成和手语的形态研究的下一步。