GA approach to invariant matching: under noise and geometric transformations

The paper proposes a new approach to solve the matching problem, in which the genetic algorithm is used as the search strategy to find the optimal matching between the features of the object and those of the model. Feature points matching is used as an example to discuss this kind of approach. Firstly the invariant relative attributes of features are discussed. With the invariant attributes, we analyse the expression of matching points. Secondly by using the genetic algorithm´s schema we design the new approach to the invariant matching. The experiments show that the algorithm is indeed invariant to the geometric transformations and it can work in the noise case.<>