Capturing Outlines of Planar Images by Fuzzy Randomized Knot Insertion to Cubic Spline

The proposed work, in this paper, is concerned with an efficient technique of curve fitting using cubic splines. The technique has various phases including extracting outlines of images, detecting corner points from the detected outline, addition of extra knot points if needed. The last phase makes a significant contribution by making the technique automated. It uses the idea of knot insertion in a fuzzy randomized manner. The proposed algorithm is an iterative one. The algorithm proposed is computationally efficient as compared to least square approach.