Classification Based Navigation and Retrieval for Picture Archives
Sean Bechhofer, Carole Goble · 1999
Current state of the art in image retrieval and indexing doesn’t meet all the needs of users of electronic picture collections. Content based retrieval provides little support for semantic metadata , in particular descriptions of what the image contains or represents. We present the approach being taken by the STARCH project, which is using a Description Logic (DL) for semantic metadata. The structured representation of the DL can assist in providing more powerful environments for retrieval, through the support of browsing, navigation and the serendipitous discovery of information. The conceptual space can also prove useful for defining notions of similarity and semantic closeness. We illustrate these claims with a series of examples taken from our prototype system. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.