An Evolution Strategies Based Approach to Image Registration

An image registration approach based on Evolution Strategies is proposed. In image registration, an invariant reference needs to be established within each source image, which is unavailable in many cases. To solve this problem, feature configuration (which is defined as the cluster of feature vectors on an image representing homogeneous feature distribution,) is employed to describe the object inside the scene. Instead of finding the correspondence of the entire image, the spatial relationships of the feature configuration in every source image are discovered with Evolution Strategies (ES). While one approach, even this one, may not be suitable for every image domain, the ES approach has many advantages. Compared to some methods, it is computationally effective; compared to other, it is capable of discovering transformations of larger scope (e.g., greater rotation angles or translation distances etc.) The search structure we use is an ellipsoid. The results from various images prove it to be an efficient and effective method.