A Neutral Framework for Feature Definition and a Generic Algorithm for Feature Recognition

Abstract

Traditional feature recognition approaches predefine features which can be recognized in systems. Feature definitions and recognitions are characterized by the domains which the approaches are applied to. Therefore, these approaches are not generic. In order to fulfill various feature recognition requirements of different users in different domains, this paper proposes an application independent framework for form feature definition and a generic algorithm for feature recognition based on the framework. Users in different domains can define their own features following the rules of the feature definition framework according to their tasks and user-defined features can be recognized by the uniform recognition algorithm. Hence, the system is flexible and can recognize a wide range of form features in different domains. The testing results show that this approach is effective in defining and recognizing isolated features as well as interacting features.