A Study on Construction of Japanese Sign Language Electronic Dictionary using 3D Animation

本研究的目的是建立一个兼顾手语识别、动画自动生成的电子词典系统。在本研究中，我在研究期间，以三维动画建构电子辞典系统，在手语辨识部分、语音分析部分、发音产生部分、影像资料库部分的各个功能上进行了检视。1.在辨识部分，我提出了以单目摄影机模型为HMM模型，以手部形态为模型，以无限制的手势发音辨识技术。2.在语言分析部分，提出了NVGS模型，使超音段音素的语素和记法成为可能，并提出了手语的语音符号记法，用于手语的发音描述。3.在发音生成部分，提出了用规则方法对所提出的发音符号进行完全合成，使非人工信号的表示成为可能。我能够利用它 ...更多信息 4.在图像数据库部分，我为电子词典录制了图像数据库KOSIGN Ver.2，其中包括大约1，100个词汇。这个数据库考虑到我在教育中使用它，它是用两个摄像机模型从左对角线前方和前方构建的一个男人和一个女人的数据。通过多媒体环境下的交互式三维动画实现了手语电子词典系统的原型，取得了较好的效果，基本达到了本课题的研究目的和实施进度。这是在1998年1月18日星期日的NHK电视台NEWS7中介绍的关于动画自动生成的结果。并且，对于去年的调度，我做了一个从这个研究调度整体提供的结果报告。本文进一步总结了本研究成果的设备工作状态，从本研究成果中分析了手语的形式，对获取与数据处理设备的多模态接口的技术因素进行了研究，并继续进行开发。少

It is that purpose of this research builds the electronic dictionary system which considered recognition of sign language, automatic generation of animation. In this research, I examined it in each function of sign language recognition part, speech analysis part, articulation generation part, image database part on construction of electronic dictionary system by 3D animation through research period.1.In recognition part, I proposed technique of no restriction by grobal articulation recognition by using monocular camera model for HMM model, hand configuration recognition.2.In linguistic analysis part, I proposed NVGS model that made a morpheme and notation of suprasegmental phoneme possible.And I proposed phonetic sign notation for articulation description of sign language.3.In articulation generation part, I proposed complete synthesis by rule method from articulation notation symbol proposed and made representation of a non-manual signal possible. I was able to have been to utilize it … More on Internet furthermore.4.I recorded image database KOSIGN Ver.2 for the electronic dictionary which consisted for about 1,100 vocabulary in image database part. This database considered that I used it in education, and it was constructed data of a man and a woman with two camera model from the left diagonal front and from the front. And this integrated function of each part and built prototype of sign language electronic dictionary system by interactive 3D animation in multi-media environment.A good result was provided as above, and purpose of this research and enforcement scheduling were almost achieved. And this were introduced with NEWS7 of NHK television on animation automatic generation among result on Sunday, January 18,1998. And, for the scheduling last year, I made report of result provided from this research scheduling whole. It is equipment working state to summarize result of this research in an article furthermore.Toward acquisition of factor technique of multimodal interface with data processing equipment by analyzing modality of sign language from this research result, I researched it, and continuation felt importance to let develop. Less