SAP: A Model for the Syntactic Analysis of Pictures
Abstract
A syntax-directed model is presented which is able to recognize and generate two-dimensional pictures while allowing a high degree of man/machine interaction. Starting with a field of points representing the picture, a string of symbols providing a structural description of the picture is produced by the syntactic component. The structural description, composed of higher level primitives (e.g., geometric symbols such as triangles and rectangles) and syntactic relations which exist between the primitives, is operated upon by the semantic component to provide a semantic interpretation for the picture. The syntactic component consists of a lexicon, a modified context-sensitive phrase structure grammar, and a set of transformation rules. The semantic component consists of a set of heuristics to abstract the picture and a modified context- sensitive phrase structure grammar which allows contextual restrictions to be applied to combinations of constituents existing at different levels of the syntax tree which syntactically describes the picture. Various aspects of the model have been programmed on the LINC (a small digital computer), the IBM 360/ 50, and the IBM 7072.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1968
- Accession Number
- AD0771739
Entities
People
- Armond David Inselberg
- Raymond M. Kline
Organizations
- University of Washington