Image Information Service with Integrated Content-based Retrieval Facilities

(1)基于内容的图像检索算法：提出了图形特征参数，在生理水平上模拟人类视觉感知过程的神经网络。我们还建立了相邻细胞之间相互作用的通用模型，如局部范围的侧抑制，以及全局范围内的整体视觉细胞，如光适应。(2)互联网上的图像数据库服务器：我们开发了数据库代理机制。它管理对不同标签结构的数据库服务器的并行访问，并通过单一用户界面将来自它们的响应整合为单一格式。(3)集成图形用户界面：通过界面代理机制开发了直观交互的用户界面。它应用基于用户感性模型的信息过滤，在认知层面和视觉层面为用户提供友好的交互。

(1) Content-based Retrieval Algorithms for Various Kinds of Images : We developed graphical feature parameters to simulate the neural network of human visual perception process at physiological level. We also developed generalized model for mutual interaction among neighboring cells such as lateral inhibition at local scope as well as whole visual cells such as light adaptation at global scope.(2) Image Database Server on the Internet : We developed database agent mechanism. It manages parallel access to the database servers in different tag structures and integrates the responses from them into the single format through the single user interface.(3) Integrated Graphical User Interface : We developed user interfaces for intuitive interaction by interface agent mechanism. It applies information filtering based on the Kansei model of the user to provide user friendly interaction at cognition level as well as visual level.