On non-uniform rational B-splines surface neural networks

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

A novel bi-variate non-uniform rational B-splines (NURBS) surface neural network consisting of four hidden layers is proposed in this paper. The blending functions are selected as the activation functions for the neurons in one of the hidden layers, instead of the commonly used sigmoid functions. With mathematical derivations, it is easy to find that the mathematical expression of the output of the proposed neural network is exactly the same as the NURBS surface. Since a set of 2-D gray scale image data can be considered as a 3-D surface, therefore the proposed NURBS surface neural network can be applied to deal with image processing problems. Two experiments, concerning image compression and corrupted image restoration, are conducted to demonstrate the feasibility of the proposed approach.

Original languageEnglish
Pages (from-to)1-15
Number of pages15
JournalNeural Processing Letters
Volume28
Issue number1
DOIs
Publication statusPublished - 2008 Aug 1

Fingerprint

Data Compression
Sigmoid Colon
Splines
Neural networks
Neurons
Image compression
Image reconstruction
Image processing
Chemical activation
Experiments

All Science Journal Classification (ASJC) codes

  • Software
  • Neuroscience(all)
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this

@article{2961f213e45f4445adc1faceab83403f,
title = "On non-uniform rational B-splines surface neural networks",
abstract = "A novel bi-variate non-uniform rational B-splines (NURBS) surface neural network consisting of four hidden layers is proposed in this paper. The blending functions are selected as the activation functions for the neurons in one of the hidden layers, instead of the commonly used sigmoid functions. With mathematical derivations, it is easy to find that the mathematical expression of the output of the proposed neural network is exactly the same as the NURBS surface. Since a set of 2-D gray scale image data can be considered as a 3-D surface, therefore the proposed NURBS surface neural network can be applied to deal with image processing problems. Two experiments, concerning image compression and corrupted image restoration, are conducted to demonstrate the feasibility of the proposed approach.",
author = "Ming-Yang Cheng and Wu, {Hung Wen} and Wen-Yu Su",
year = "2008",
month = "8",
day = "1",
doi = "10.1007/s11063-008-9078-9",
language = "English",
volume = "28",
pages = "1--15",
journal = "Neural Processing Letters",
issn = "1370-4621",
publisher = "Springer Netherlands",
number = "1",

}

On non-uniform rational B-splines surface neural networks. / Cheng, Ming-Yang; Wu, Hung Wen; Su, Wen-Yu.

In: Neural Processing Letters, Vol. 28, No. 1, 01.08.2008, p. 1-15.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On non-uniform rational B-splines surface neural networks

AU - Cheng, Ming-Yang

AU - Wu, Hung Wen

AU - Su, Wen-Yu

PY - 2008/8/1

Y1 - 2008/8/1

N2 - A novel bi-variate non-uniform rational B-splines (NURBS) surface neural network consisting of four hidden layers is proposed in this paper. The blending functions are selected as the activation functions for the neurons in one of the hidden layers, instead of the commonly used sigmoid functions. With mathematical derivations, it is easy to find that the mathematical expression of the output of the proposed neural network is exactly the same as the NURBS surface. Since a set of 2-D gray scale image data can be considered as a 3-D surface, therefore the proposed NURBS surface neural network can be applied to deal with image processing problems. Two experiments, concerning image compression and corrupted image restoration, are conducted to demonstrate the feasibility of the proposed approach.

AB - A novel bi-variate non-uniform rational B-splines (NURBS) surface neural network consisting of four hidden layers is proposed in this paper. The blending functions are selected as the activation functions for the neurons in one of the hidden layers, instead of the commonly used sigmoid functions. With mathematical derivations, it is easy to find that the mathematical expression of the output of the proposed neural network is exactly the same as the NURBS surface. Since a set of 2-D gray scale image data can be considered as a 3-D surface, therefore the proposed NURBS surface neural network can be applied to deal with image processing problems. Two experiments, concerning image compression and corrupted image restoration, are conducted to demonstrate the feasibility of the proposed approach.

UR - http://www.scopus.com/inward/record.url?scp=42949172078&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=42949172078&partnerID=8YFLogxK

U2 - 10.1007/s11063-008-9078-9

DO - 10.1007/s11063-008-9078-9

M3 - Article

VL - 28

SP - 1

EP - 15

JO - Neural Processing Letters

JF - Neural Processing Letters

SN - 1370-4621

IS - 1

ER -