Construction of Web Camera Searching Environment Using the Next Generation Web

近年来，作为下一代Web的语义Web引起了世界各地研究人员的极大关注，与之相关的研究成果在ISWC（International Semantic Web Conference）等各种国际会议上被积极报道。在日本，研究成果也可以在各种学术团体中找到，如Web Intelligence and Interaction society。虽然语义Web已经引起了研究者的极大关注，但是从用户的角度来看，语义Web还没有被人们所熟知，还只是一个正在进行的研究课题，我们认为这是因为目前对语义Web的研究很少关注其应用，而过于关注其基础设施。因此，本研究将研究重点放在语义网的应用上，具体构建了一个基于语义网的网络摄像头智能检索环境（2007年），我们实现了一个发布Web摄像机的平台，包括用于标记Web摄像机的元数据模式，元数据的表示方法，以及网络摄像头和元数据的发布工具。2008年，我们使用该平台构建了一个网络摄像头搜索环境。然后，在2009年，研究的最后一年，我们研究了我们的网络摄像头搜索环境的性能进行了广泛的实验，并评估的精度，召回率，和F-测量的环境。

Recently, as the next generation of Web, the Semantic Web has attracted considerable attention of researchers around the world, and research results related to it have been being actively reported at various international conferences such as ISWC (International Semantic Web Conference). In Japan, research results could also be found in various academic societies like the Web Intelligence and Interaction society. Although it has attracted considerable attention of the researchers, from viewpoint of the users, we have to say the Semantic Web has not been well known and is still no more than a research subject in progress.We think that this is because the present research on Semantic Web pay little attention to the application; it pays too much attention to the infrastructure. Thus, in this research, we focus our attention on the application of Semantic Web, and in particular, we construct practically a Web Camera Searching Environment in which Web cameras can be intellectually retrieved over the Semantic Web.Given such a background of research, in the first year (year 2007), we implemented a platform for publishing Web cameras, including a schema of metadata for tagging Web cameras, a presentation method of metadata, and a publishing tool of Web Camera and Metadata. In 2008, using the platform, we constructed a Web Camera Searching Environment. Then in 2009, the last year of the research, we studied the performance of our Web Camera Searching Environment by conducting extensive experiments, and evaluating the precision, recall rate, and F-measure of the environment.