Abstract:
A new method, which is employing Adaptive Genetic Algorithms (AGA), is developed for the detection and parameter evaluation of elliptic ring-shaped clusters. The algorithm is based on optimization of an objective function employing AGA, through which we can not only speed up the clustering, but avoid being trapped into minimum. The algorithm is also validated through testing various forms of synthetic data, such as multiple intersecting ellipses, concentric ellipses, incomplete ellipses, and etc.