Feature Extraction: Issues, New Features, and Symbolic Representation

Feature extraction is an important part of object model acquisition and object recognition systems. Global features describing properties of whole objects,or local features denoting the constituent parts of objects and their relationships may be used. When a model acquisition or object recognition system requires symbolic input,the features should be represented in symbolic form. Global feature extraction is well-known and oft-reported. This paper discusses the issues involved in the extraction of local features, and presents a method to represent them in symbolic form. Some novel features, specifically between two circular arcs, and a line and a circular arc, are also presented.