Symbolic Pruning in a Structural Approach to Engineering Drawing Analysis

Interpretation of paper drawings has received a good deal of attention over the last decade. Progress has also been made in related areas such as direct interpretation of human drawings (HCI), search and indexing of graphics databases, and knowledge representation in the domain of graphics and drawing understanding. One of the most interesting applications in this domain is the analysis of semantics in engineering drawings. We propose a structural modeling approach combined with a nondeterministic agent system to produce interpretations of scanned images of CAD drawings. We allow a broad set of thresholds during the image analysis and show how this large search space can be ef£ciently pruned by taking advantage of constraints from the model or speci£c application.