A knowledge structuring technique for efficient classification in multiple scene analysis

The growing importance of automatic scene analysis is verified in many fields like robotics in factory automation, medical applications, monitoring and control, (environmental control is an important emerging discipline), and others. This considerable growth of interest has created an increasing need for both theoretical methods and experimental software and hardware. The processing steps for image processing and pattern recognition are well established and a lot of algorithmic solutions are nowadays available. The main steps typically are: frame preprocessing, frame segmentation, object classification and recognition. In this paper, the authors propose an approach which calls for generalization and flexibility in the system design and object representation, based on object oriented programming, (OOP), paradigm concepts for knowledge structuring