The research on Modeling the Attitude for Art using Appreciation Robot

1998年，我们通过安装CCD摄像机开发了一台遥控博物馆机器人的原型。我们把遥控机器人连上了互联网。美国卡内基梅隆大学的网络远程控制实验取得了成功。在这项研究中，我们计划讨论我们对图像的识别取决于我们如何通过文字识别图像。通过网络研究，将原始图像和视觉效果图像呈现给家庭中的受试者。受试者对图像词作出反应，根据图像词进行联想。由此，通过结构化分析，我们可以得到基于隐喻(1)对象、(2)媒体、(3)原因和(4)情感四个结构元素的用户图像识别的感性信息过程。观察物体的态度可以概括为：第一类：通过将“伊森”与物体的刺激联系起来进行假设推理的类型。第二组：通过客体刺激进行“情绪表达”的类型。第三组：通过将“媒介”与物体的刺激联系起来做出合乎逻辑的结论的类型。组4：将客体的刺激表达为“客体”的类型。第5组：表达态度的类型，包括上述所有类别。从原始图像中提取往往受到印记在图像中的上下文质量的影响，以及对图像的评价往往受到个人品味差异的影响。

In 1998, we have developed a prototype of the remote control museum robot by mounting the CCD camera. We connected remote contoled robot with internet. The experiment of remote controlling through internet was done successfully from the University of Carnegy Mellon in USA.In the present study, we plan to discuss what we recognize as an image depends on how we recognize images by words. Original images and visual effected images were presented to subjects in a home-age through network research. Subjects responded to image words making associations from the image. From this, by structured analysis, we could obtain a user' KANSEI information process of image recognition based on 4 structural elements of metaphors (1)object, (2) media, (3)reason, and (4) emotion. The attitude observing an object can be summarized as follows.Group 1 : The type that makes hypothetical reasoning by associating "eason" with the stimulation from the object. Group 2 : The type that makes "emotional expression" with the stimulation from the object. Group 3 : The type that makes logical conclusion by associating "medium" with the stimulation from the object. Group 4 : The type that expresses the stimulation from the object as "object" as it is seen. Group 5 : The type that expresses attitude including all above categories.The extraction from original images often tends to be affected by the quality of the context imprinted in the images and the evaluation on the images often done by the difference of taste of the person.