Clustering of ellipses based on their distinctiveness: An aid to ellipse detection algorithms

Many ellipse detection algorithms produce multiple elliptic hypotheses corresponding to a single elliptic object. Thus, it is needed to identify similar ellipses that possibly belong to the same object and cluster them as a single object. This will reduce the computational and memory requirement for further higher level processing of ellipse detection algorithms. Here, we present a method better than Basca´s method to detect similar ellipses and cluster them. This method is independent of the ellipse detection and hypotheses generation scheme. The presented method is simple, intuitive, requires few computations, and is applicable to wide range of scenarios.