A Deep Neural Network for Knot Placement in B-spline Approximation

TL;DR

This paper presents a deep neural network approach to automatically determine the optimal number and positions of knots in B-spline approximation, improving efficiency and accuracy over traditional methods.

Contribution

Introduces a novel deep neural network framework that models knot placement as a mapping, enabling automatic and efficient optimization of knot number and positions.

Findings

The neural network effectively approximates optimal knot placement.

The method outperforms traditional algorithms in accuracy and efficiency.

Numerical examples demonstrate the superiority of the proposed approach.

Abstract

Automatically determining knot number and positions is a fundamental and challenging problem in B-spline approximation. In this paper, the knot placement is abstracted as a mapping from initial knots to the optimal knots. We innovatively introduce a deep neural network solver to approximate the mapping. The neural network is composed of several subnetworks. Each subnetwork is designed to approximate the optimal knot positions in the case of fixed knot number. All the subnetworks are stacked together to find the optimal knots (including knot number and knot positions) within some given tolerance. Owing to the powerful approximation capabilities, as well as mature algorithms developed in deep learning, the proposed method can effectively and efficiently find the optimal knot number and knot positions. Several numerical examples are demonstrated to show the superiority of our method.