A Modified Complete Spline Interpolation and Exponential Parameterization - Computer Information Systems and Industrial Management Access content directly
Conference Papers Year : 2015

A Modified Complete Spline Interpolation and Exponential Parameterization

Abstract

In this paper a modified complete spline interpolation based on reduced data is examined in the context of trajectory approximation. Reduced data constitute an ordered collection of interpolation points in arbitrary Euclidean space, stripped from the corresponding interpolation knots. The exponential parameterization (controlled by λε [0, 1]) compensates the above loss of information and provides specific scheme to approximate the distribution of the missing knots. This approach is commonly used in computer graphics or computer vision in curve modeling and image segmentation or in biometrics for feature extraction. The numerical verification of asymptotic orders α(λ) in trajectory estimation by modified complete spline interpolation is performed here for regular curves sampled more-or-less uniformly with the missing knots parameterized according to exponential parameterization. Our approach is equally applicable to either sparse or dense data. The numerical experiments confirm a slow linear convergence orders α(λ) = 1 holding for all λε [0, 1) and a quartic one α(1) = 4 once modified complete spline is used. The paper closes with an example of medical image segmentation.
Fichier principal
Vignette du fichier
978-3-319-24369-6_8_Chapter.pdf (536.64 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01444508 , version 1 (24-01-2017)

Licence

Attribution

Identifiers

Cite

Ryszard Kozera, Lyle Noakes, Magdalena Wilkołazka. A Modified Complete Spline Interpolation and Exponential Parameterization. 14th Computer Information Systems and Industrial Management (CISIM), Sep 2015, Warsaw, Poland. pp.98-110, ⟨10.1007/978-3-319-24369-6_8⟩. ⟨hal-01444508⟩
186 View
188 Download

Altmetric

Share

Gmail Facebook X LinkedIn More