Feature extraction from boundary models of three-dimensional objects

Abstract

An algorithm for extracting certain classes of form features from a relational boundary model of an object, called the generalized edge-face graph (GEFG), is described. The GEFG provides a face-based topological description of the object boundary and encodes the minimum number of relations needed in the recognition process. The feature identification and classification are based on the analysis of the connectivity properties of the edge-face graph associated with the GEFG and on some geometric considerations. The result is a hierarchical graph decomposition of the object boundary into components representing form features

Publication
Pattern Analysis and Machine Intelligence, IEEE Transactions on