Handwritten curve approximation by a Bézier curve with featured points

Abstract

Approximating a hand-written curve by a Bézier curve can be performed by two main processes: (1) the selection of the significant points on the curve and (2) the curve approximation. This paper focuses on finding a set of featured points, which is an important component for curve fitting into a Bézier curve. This new approach is an extension of the authors' previous work by reducing the weakness of point sampling. In performance evaluation, a new point-sampling algorithm is used to compare with the previous methods which are the results from the approximation of handwritten curves by Chebyshev polynomials.