FLEXIBLE MULTIMEDIA SYSTEMS BY KANSEI INFORMATION PROCESSING

在分布式多媒体数据库系统中，数据库对象由视频、音频、图形和文本数据组织，存储在广域网上不同位置的多个数据库服务器中，这对各种应用领域的设计人员使用多媒体数据库进行CAD处理，尤其需要智能和灵活的检索能力。在本研究中，我们引入了一种基于感知信息处理的分布式设计图像数据库功能的灵活超媒体系统，即感性信息方法，它允许每个用户充分利用他或她的创造力和原创性，以及他或她的风格和对美的敏感性。知识代理利用知识库，从用户提供的感性词对象到分布在互联网上的多媒体数据库中存储的合适的多媒体对象之间建立链接。然后将这些感性查询词转换为等效的颜色值和设计模式。色彩值和图案是所有存储的设计图像的表征手段。知识代理还使用每个用户的个人“用户模型”，将单个用户的主观感性（唯一的、主观的感性词使用）查询转换为客观感性词。为了研究我们建议的感性信息处理方法的实现，我们开发了一个纺织品设计图像数据库，它可以帮助设计师在不同地点、合作或单独工作时使用服装CAD系统。通过该设计艺术系大学生对原型系统的功能和性能评价，验证了所提出的感知检索方法在分布式设计图像数据库系统中的实用性。

In distributed multimedia database systems where database objects are organized by vide, audio, graphics and text data stored in multiple numbers of databaase servers on different loations over a wide area network, intelligent and flexible retrieval capability is required particularly for designeres in various application fields who user a multimedia database for their CAD process.In this research, we introduce a flexible hypermedia system for distributed design image databases capabilities based on perceptional information processing, so called Kansei information method, which allows each user to make the best use of his or her creativity and originality and his or her style and sensitvity to beauty. The knowledge agent, utillzing the knowledge-base, creates links from kansei word objects provided by the user to suitable multimedia objects among those stored on multimedia database distributed across the Internet. Those kansei query words are then converted to equivalent color values and design patterns. Color value and design pattern are the means by which all stored design image are characterized. The knowledge agent also performs query conversion of individual user's subjective kansei (unique, subjective use of kansei word) into objective kansei words using each user's individual "user model." In order to investigate the realization of our suggested Kansei information process method, we have developed a textile design image database which aids designers using apparel CAD systems in different locations, collaboratin or working separately. Through the functional and performance cvaluation of the prototype system by the design art university student, the usefulness of the suggested perceptional retrieving method for distributed design image database syste was verlfled.