Fourier-Informed Knot Placement Schemes for B-Spline Approximation

by   David Lenz, et al.

Fitting B-splines to discrete data is especially challenging when the given data contain noise, jumps, or corners. Here, we describe how periodic data sets with these features can be efficiently and robustly approximated with B-splines by analyzing the Fourier spectrum of the data. Our method uses a collection of spectral filters to produce different indicator functions that guide effective knot placement. In particular, we describe how spectral filters can be used to compute high-order derivatives, smoothed versions of noisy data, and the locations of jump discontinuities. Our knot placement method can combine one or more of these indicators to place knots that align with the qualitative features of the data, leading to accurate B-spline approximations without needing many knots. The method we introduce is direct and does not require any intermediate B-spline fitting before choosing the final knot vector. Aside from a fast Fourier transform to transfer to and from Fourier space, the method runs in linear time with very little communication. The method is applied to several test cases in one and two dimensions, including data sets with jump discontinuities and noise. These tests show that the method can fit discontinuous data without spurious oscillations and remains accurate in the presence of noise.



There are no comments yet.


page 1

page 2

page 3

page 4


Adaptive spline fitting with particle swarm optimization

In fitting data with a spline, finding the optimal placement of knots ca...

Reconstruction of piecewise-smooth multivariate functions from Fourier data

In some applications, one is interested in reconstructing a function f f...

A space-time discretization of a nonlinear peridynamic model on a 2D lamina

Peridynamics is a nonlocal theory for dynamic fracture analysis consisti...

Statistics of the Nonlinear Discrete Spectrum of a Noisy Pulse

In the presence of additive Gaussian noise, the statistics of the nonlin...

A Fast Template Periodogram for Detecting Non-sinusoidal Fixed-shape Signals in Irregularly Sampled Time Series

Astrophysical time series often contain periodic signals. The large and ...

Fourier-based and Rational Graph Filters for Spectral Processing

Data are represented as graphs in a wide range of applications, such as ...

IsoGeometric Approximations for Electromagnetic Problems in Axisymmetric Domains

We propose a numerical method for the solution of electromagnetic proble...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.