Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
High-Fidelity Image Interpolation using Radial Basis Function Neural Networks
 

Summary: High-Fidelity Image Interpolation using Radial Basis
Function Neural Networks
Abstract
Iiiiage iriierpolatioii using radial basis f'uiict ion
(RBF)neural iietworlis is accorilplislletl. 111t hi.
work the KBF net,u:ork is first trained Tvitli tile
gil-eniiiiage, satisfying the constraint of' tlic gra\-
valiie at, each pixel. With the desired magnifica-
tion ratio. each pixel is t,lien di\-icled iiito subpix-
els. The subpisel gra.y values are calculated using
the trained network. Two dimensional Gaussian
hasis functions are used as the iieuroiis in the
hidden layer.
1 Introduction
Iiriage interpolation deals with the task of ieii-
dering a high-resolution image from a lower re+
oliitioii one. Conventional interpolatioii tech-
riiques are linear, bilinear. spline. and the neai-
est neighbor interpolation [l.2, 3 . -I]. *4ll these
irietlrods aim at reducing the mean-squared el-

  

Source: Ahmed, Farid - Department of Electrical Engineering and Computer Science, Catholic University of America

 

Collections: Engineering