High level scene interpretation using fuzzy belief

Abstract

In this paper we present an image understanding system using fuzzy sets. This system is based on a symbolic object-oriented image interpretation system (SOO-PIN) we developed previously. It is known that in many image analysis and understanding applications, objects are not well-defined and are engaged in dynamic activities, which in most cases can only be described vaguely. Using fuzzy sets we are able to capture subtle variations and manage uncertainty properly. We demonstrate the effectiveness of our system with complex traffic scenes.