Fine Grained Semantics of Images in Digital Museum
Zhang Gaofeng, Tongshan Liu, Shaobo Liu, Zhou Qingguo · 2010
There are many murals in Mogao Grottoes. Most of murals describe specific stories involving religion, history, culture and so on. These stories are too specializing to understanding for nonprofessional visitors. To improve visitors understand the content of these murals, a novel technology named Fined Grained Semantics of Images(FGSI) is studied and applied in Dunhuang Studies Digital Museum. With the technology, more fine-grained and human understandable semantics is put to digital images of murals. In other words, take the semantic unit in image as little as possible. For instance, a specific person or animal in the image may be taken as a role in corresponding story. So the person or animal will be taken as an independent semantic unit and own the necessary semantic attributes, such as name etc. In virtual exhibitions, the image and corresponding text will be put into one page. With FGSI, semantics of image and text are associated with each other. When semantic unit of image get the focus such as cursor is on it, the corresponding text will be highlighted such that visitors can conveniently see who and what is doing in murals.