DocumentCode :
3500860
Title :
An optimal construction and training of second order RBF network for approximation and illumination invariant image segmentation
Author :
Cai, Xun ; Tyagi, Kanishka ; Manry, Michael T.
Author_Institution :
Sch. Of Comput. Sci. & Technol., Shandong Univ., Jinan, China
fYear :
2011
fDate :
July 31 2011-Aug. 5 2011
Firstpage :
3120
Lastpage :
3126
Abstract :
In this paper, we proposed an hybrid optimal radial-basis function (RBF) neural network for approximation and illumination invariant image segmentation. Unlike other RBF learning algorithms, the proposed paradigm introduces a new way to train RBF models by using optimal learning factors (OLFs) to train the network parameters, i.e. spread parameter, kernel vector and a weighted distance measure (DM) factor to calculate the activation function. An efficient second order Newton´s algorithm is proposed for obtaining multiple OLF´s (MOLF) for the network parameters. The weights connected to the output layer are trained by a supervised-learning algorithm based on orthogonal least square (OLS). The error obtained is then back-propagated to tune the RBF parameters. By applying RBF network for approximation on some real-life datasets and classification to reduce illumination effects of image segmentation, the results show that the proposed RBF neural network has fast convergence rates combining with low computational time cost, allowing it a good choice for real-life application such as image segmentation.
Keywords :
Newton method; backpropagation; image classification; image segmentation; least squares approximations; radial basis function networks; RBF learning algorithm; RBF network training; activation function; backpropagation; hybrid optimal radial-basis function neural network; illumination effect reduction; illumination invariant image segmentation; image classification; kernel vector; network parameter training; optimal RBF network construction; optimal learning factors; orthogonal least square; second order Newton algorithm; second order RBF network; spread parameter; supervised learning algorithm; weighted distance measure factor; Classification algorithms; Delta modulation; Image color analysis; Image segmentation; Radial basis function networks; Support vector machine classification; Training;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks (IJCNN), The 2011 International Joint Conference on
Conference_Location :
San Jose, CA
ISSN :
2161-4393
Print_ISBN :
978-1-4244-9635-8
Type :
conf
DOI :
10.1109/IJCNN.2011.6033634
Filename :
6033634
Link To Document :
بازگشت