Object Area-based Method for Elliptic and CircularFit of a Two-tone Image

Abstract

Circular or elliptic fit of an object is  important  in target detection, shape analysis, and biomedical imageanalysis problems. Here, the problems of fitting circle or ellipse to an object in 2-D are considered. At first,the problems is converted to quadratic equation of single unknown by some constraints. Then its solutionsfor all the border points of the object are found and averaged. The major and minor axes of ellipse are presentedby least sum perpendicular distance of all points of the object. The other unknowns are found  using theequations of constraints. In the proposed method, the main constraint used for the circular and elliptic fit isthat the area of the fitting circle or ellipse is equal to the area of the object to be fitted. The approach appearsto be less sensitive to the object border noise and is computationally attractive. Some examples are presentedto show the effectiveness of the approach. A measure of degree of circularity, ellipticity, etc in fuzzy settheoretic framework is also proposed.

Defence Science Journal, 2008, 58(6), pp.710-714, DOI:http://dx.doi.org/10.14429/dsj.58.1696