Title :
Approximating Handwritten Curve by Using Progressive-Iterative Approximation
Author :
Nuntawisuttiwong, Taweechai ; Dejdumrong, Natasha
Author_Institution :
Dept. of Comput. Eng., King Mongkut´s Univ. of Technol., Bangkok, Thailand
Abstract :
In this paper, a given handwritten curve will be approximately converted into a Bezier curve. First, a number of sampled points on the given curve are systematically chosen to guarantee that the Bezier curve can be adequately modeled without the loss of curve characteristics. These points are considered to be the Bezier control points. The Progressive-Iterative Approximation (PIA) is then adopted to construct the interpolating curve. Employing PIA algorithm, Bernstein polynomials are selected to be a blending function of PIA so that this interpolating curve will develop a Bezier curve. This technique will avoid the ill-conditioned circumstance because of using Newton/Lagrange polynomials in curve fitting.
Keywords :
Newton method; curve fitting; handwritten character recognition; interpolation; polynomial approximation; Bernstein polynomials; Bezier control; Bezier curve; Newton-Lagrange polynomials; PIA algorithm; curve characteristics; handwritten curve approximation; interpolating curve; progressive-iterative approximation; Approximation algorithms; Computer graphics; Interpolation; Polynomials; Shape; Vectors; Approximated Handwritten; Interpolation; Progressive-Iterative Approximation;
Conference_Titel :
Computer Graphics, Imaging and Visualization (CGIV), 2013 10th International Conference
Conference_Location :
Macau
DOI :
10.1109/CGIV.2013.15