Geometrical constraints for object recognition using genetic algorithms

In this paper we describe how different types of constraints can affect the performance of genetic algorithms (GAs). The success of a GA application depends on efficient constraints to provide a clear direction for GA search. Our application integrates geometrical constraints with a GA to guide the pattern matching process for image registration and object location. Two types of constraints, namely, local and global constraints are considered Although both types of constraints are useful in constraining the pattern matching search space, the former type is less effective than the latter type. Intuitively one would expect that the combination of both types of constraints should be more powerful than each of them used alone. Yet our experimental result proves to the contrary. We describe the whole process of integrating geometrical constraints with a GA for pattern matching and analyse the results