This paper presents a new approach for precision estimation for algebraic ellipse fitting based on combined least squares method. Our approach is based on coordinate description of the ellipse geometry to determine the error distances of the fitting method. Since it is an effective fitting algorithm the well-known Direct Ellipse Fitting method was selected as an algebraic method for precision estimation. Once an ellipse fitted to the given data points, algebraic distance residuals for each data point and fitting accuracy can be computed. Generally, the adopted approach has revealed geometrical aspect of precision estimation for algebraic ellipse fitting. The experimental results revealed that our approach might be a good choice for precision estimation of the ellipse fitting method.
Ahn S J, Rauh W, Warnecke H S 2001: Pattern Recognition, 34, 2283–2303.
Warnecke H. S., '' (2001) 34Pattern Recognition: 2283-2303.
Warnecke H. S.Pattern Recognition20013422832303)| false
Cui Y, Weng J, Reynolds H 1996: Pattern Recognition Lett., 17, 309–316.