Formal system based on fuzzy logic applied to digital image scene analysis

A new approach of the digital image scene analysis is presented. In order to provide a flexible theoretical background for digital image data acquisition, image object recombining and image scene analysis, a formal system is introduced. In order to implement a formal system in a context of digital image analysis, a formal language is also introduced. The basic idea behind it is that simple parts of complex objects can be efficiently recognized and then step by step recombined back to the complex objects. During this process of recombining, an analysis of object features and relations among them can be obtained. In order for image scene analysis become feasible, local image information should be coded in the same way as global information of an image can be reconstructed later. Digital image feature detection and object recognition can be seen as a part of scene analysis. A formalization of object classification using formal language scene analysis is given. A computer implementation of a formal system suitable for digital image analysis is presented. Simple examples of digital image scene analysis are given.