Fuzzy Mathematical Morphology and Derived Spatial Relationships
Isabelle Bloch · Studies in fuzziness and soft computing · 2000
The joint importance of mathematical morphology in image processing and of fuzzy sets for representing imprecisely defined image objects has led to an increased interest in fuzzy mathematical morphology. In this chapter, we first review the main definitions of fuzzy mathematical morphology operators and show that the most important properties of classical mathematical morphology are kept. Then we use these operators, mainly fuzzy dilation, for defining some spatial relationships between fuzzy objects: distances, adjacency and boundary, directional relative position. Such spatial fuzzy relationships are important in pattern recognition (for instance model-based structural pattern recognition). Indeed, objects can be recognized in a scene using their own characteristics, but also using their relationships to other objects of the scene. Therefore the proposed operations and relationships between fuzzy objects lead to applications in structural pattern recognition under imprecision.