Cuckoo Search Algorithm for Optimal Choice of Free Knots in B-spline Data Fitting

Fitting data points to curves commonly known as curve reconstruction a significant problem in computer aided design/manufacturing (CAD/CAM). This problem is frequently encountered in the field of reverse engineering where a free-form parametric curve (typically a B-spline) with a set of (usually a high-dimensional and noisy) data points, obtained by 3D laser scanning, has to be fitted. Although there are a number of methods to come up with this problem, until now there has not been a satisfactory general solution to the problem. In this study, the cuckoo search algorithm (CS), one of the optimization methods inspired by a bird species named cuckoo that leave their eggs in the nest of other host birds, is used to solve the problem of curve fitting. Reverse engineering is used to obtain the curve from the data points. In addition, the knot positions and number of knot are free variables of the problem in the estimation of the curve, and these parameters are randomly selected in the search space by the CS method. In this way, the curve estimate with the smallest error rate is aimed to obtain in this study. Five different functions frequently used in the literature for curve fitting are preferred in the experimental studies. In the experimental results, the original curve and the predicted curve for each function are presented comparatively, and the results obtained show that for most functions, the curves predicted by the CS method produce very similar results to the original curve