Two-dimensional images of fabricated components are interpreted by querying a novel form of model database, consisting of an object-centred attribute structure and a cellular structure organized as a series of four binary relation cued inverted lists. An experimental study of the system performance is presented and compared with several optional approaches presented elsewhere. © 1987.
|Number of pages||7|
|Journal||Image and Vision Computing|
|Publication status||Published - May 1987|
- heuristic search
- image interpretation
- knowledge-based systems