Content-based image retrieval technology is vital for effective searching and browsing in the large archive of imagery. To realize content-based image retrieval, this paper proposes the ``hierarchical model of image content elements'' as the image representation model. This model consists of five layers, where each layer defines image content elements to be extracted and indexed, and similarity metric associated with them. Image database systems are then considered under network environment, and the idea of a meta search engine for distributed image database systems is proposed to improve accessibility to various contents of satellite imagery. Finally the results of content-based image retrieval demonstrate the effectiveness and potential of content-based image database systems.

      @InProceedings{ k:wains5,
      author = {Asanobu KITAMOTO},
      title = {Toward Content-Based Satellite Image Database Systems over the Network},
      booktitle = {Proceedings of the 5th International Workshop on Academic Information Networks and Systems (WAINS)},
      pages = {31-38},
      year = 1998,
      month = 12,
      doi = {},
      note = { (in English)},
      }