SOME ALGORITHMS FOR CURVE DESIGN AND AUTOMATIC OUTLINE CAPTURING OF IMAGES

A new multipurpose curve technique has been introduced which is meant to automatically provide a fit to any ordered data in a plane. The technique is particularly economical for designing purposes as well as for the visualization of a large amount of data sets. A more flexible class of cubic functions is the basis of this technique. This class of functions involves two control parameters, to produce more flexible shapes than ordinary Bézier cubics or Hermite cubics, in each segment. These functions, together with the control parameters, are utilized to fit a design curve in an interactive way. These functions are also utilized in an optimal way to fit a design curve to the data arose from any image or a scientific phenomenon. The design curve method is highly useful to capture the outlines of images. It differs, in its methodology, from the existing techniques in the literature using Bézier cubics. The curve technique has used various ideas in its construction. These ideas include end-point interpolation, detection of characteristic points, least squares approximation. The final shape is achieved by stitching the generalized Bézier cubic pieces with GC1 smoothness. Finally, three algorithms have been proposed for various applications.