{"type":"ScholarlyArticle","title":"An Approach to Knowledge-Directed Image Analysis","authors":[{"name":"Dana H. Ballard","slug":"dana-h-ballard","given":"Dana H.","family":"Ballard","affiliation":""},{"name":"Christopher M. Brown","slug":"christopher-m-brown","given":"Christopher M.","family":"Brown","affiliation":""},{"name":"Jay M. Feldman","slug":"jay-m-feldman","given":"Jay M.","family":"Feldman","affiliation":""}],"abstract":"A vision system is described which uses a semantic network model and a distributed control structure to accomplish the image analysis process. In our view, the process of analyzing an image involves directing attention through geometrical constraints, and the application of procedures attached to different conceptual nodes in the network. Several procedures are typically provided for each conceptual node because, e.g., there may be a need to find the same object in the image in very different contexts. These procedures each have a description containing performance measures and pre- and post-conditions. The descriptions are examined by an executive procedure to determine the procedure that is most useful in a context. The result of analyzing a particular image is a sketchmap, which is an instantiation of the relevant part of the larger generic network model. The current system attempts to resolve several of the fundamental problems in vision. The semantic network facilitates the interplay between geometric and other relational constraints which are used to direct and limit search. The use of multiple attached procedures allows the modelling of objects under radically different conditions and levels of detail. This tends to make the system robust in that an object which could not be located initially may be found later when the knowledge about the image has increased.","doi":null,"url":"https://ijcai.org/proceedings/1977/159","pdf":"https://ijcai.org/proceedings/1977/0159.pdf","markdown":"https://ijcai.org/proceedings/1977/159.md","bibtex":"https://ijcai.org/proceedings/1977/bibtex/159","volume":{"slug":"1977","title":"Proceedings of the Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 1977","url":"https://ijcai.org/proceedings/1977/","year":1977,"location":"Cambridge, MA, USA","dates":"","editors":"Raj Reddy","isbn":"","publisher":"International Joint Conferences on Artificial Intelligence","doi_prefix":""},"section":"Vision","track":"Vision","number":159,"pages":{"first":664,"last":669},"keywords":[],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}