A technique for the selection of kernel-function parameters in RBF neural networks for classification of remote-sensing images

@article{Bruzzone1999ATF,
  title={A technique for the selection of kernel-function parameters in RBF neural networks for classification of remote-sensing images},
  author={Lorenzo Bruzzone and Diego Fern{\'a}ndez-Prieto},
  journal={IEEE Trans. Geosci. Remote. Sens.},
  year={1999},
  volume={37},
  pages={1179-1184}
}
A supervised technique for training radial basis function (RBF) neural network classifiers is proposed. Such a technique, unlike traditional ones, considers the class memberships of training samples to select the centers and widths of the kernel functions associated with the hidden neurons of an RBF network. The result is twofold: a significant reduction in the overall classification error made by the classifier and a more stable behavior of the classification error versus variations in both… 

Figures and Tables from this paper

A radial basis function neural network based on artificial immune systems for remote sensing image classification

  • Q. YanY. Zhong
  • Computer Science
    International Conference on Earth Observation for Global Changes
  • 2008
The artificial immune network (aiNet) model, a new computational intelligence based on artificial immune networks (AIN), is applied to obtain appropriate centers for remote sensing image classification and it is shown that the aiNet-RBF NN algorithm outperform other algorithms and provides an effective option for remoteensing image classification.

A comparative study of radial basis function neural networks and wavelet neural networks in classification of remotely sensed data

This study compares radial basis function neural networks (RBFNN) and wavelet neural networks for multispectral image classification and an attempt is made to study the performance of the RBFNN with the centers and widths chosen using the classical k-means clustering.

RBF two-stage learning networks exploiting supervised data in the selection of hidden unit parameters: an application to SAR data classification

  • A. BaraldiP. BlondaG. SatalinoA. D’AddabboC. Tarantino
  • Computer Science
    IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Proceedings (Cat. No.00CH37120)
  • 2000
In the field of remotely sensed image classification, the authors compare a traditional RBFTwo-stage hybrid learning procedure with an RBF two-stage learning technique exploiting labeled data to adapt hidden unit parameters.

Hyperspectral image classifications and feature selection

  • M. Pal
  • Computer Science, Environmental Science
  • 2020

Support vector machines for classification of hyperspectral remote-sensing images

  • F. MelganiL. Bruzzone
  • Environmental Science, Mathematics
    IEEE International Geoscience and Remote Sensing Symposium
  • 2002
This paper investigates the effectiveness of SVMs in terms of classification accuracy, computational time and stability to parameter setting, and suggests them as a promising tool to classify hyperspectral remote-sensing images.

HYPERSPECTRAL IMAGE CLASSIFICATION USING SUPPORT VECTOR MACHINES: A COMPARISON WITH DECISION TREE AND NEURAL NETWORK CLASSIFIERS

The results of applying SVMs for a 16-class classification of an AVIRIS image are presented and the SVM classifier was more computationally efficient than the RBF classifier for hyperspectral image classification.

Classification of multi-sensor remote sensing images using self-organizing feature maps and radial basis function networks

  • C. ChenB. Shrestha
  • Environmental Science, Computer Science
    IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Proceedings (Cat. No.00CH37120)
  • 2000
The authors propose the use of self-organizing feature map (SOFM) neural networks to evaluate the centers for the hidden neurons in RBF neural networks for pixel classification to demonstrate the advantage of multi-sensor remote sensing using neural networks.

RBF network with two-stage supervised learning: an application

  • P. BlondaA. BaraldiA. D’AddabboC. Tarantino
  • Computer Science
    2002 Annual Meeting of the North American Fuzzy Information Processing Society Proceedings. NAFIPS-FLINT 2002 (Cat. No. 02TH8622)
  • 2002
The results indicate that the supervised approach performs better than the traditional approach when the number of hidden unit is the same and seems more stable to changes in the numberof hidden units.

A KNOWLEDGE-INTEGRATED RBF NETWORK FOR REMOTE SENSING CLASSIFICATION

A knowledge-integrated RBF model that combines the power of approximation in high -dimensional space of the RBF network and the logic reasoning of rule -based inference is proposed in the present study and experimental results show that the proposed model is more accurate, faster in training, simple in structure, and more interpretable.
...

References

SHOWING 1-10 OF 14 REFERENCES

An experimental comparison of neural and statistical non-parametric algorithms for supervised classification of remote-sensing images

Strategies and best practice for neural network image classification

It is concluded that a vast body of accumulated experience is now available, and that neural networks can be used reliably and with much confidence for routine operational requirements in remote sensing.

Neural Network Approaches Versus Statistical Methods in Classification of Multisource Remote Sensing Data

Neural network learning procedures and statistical classificaiton methods are applied and compared empirically in classification of multisource remote sensing and geographic data. Statistical

Neural networks for pattern recognition

This is the first comprehensive treatment of feed-forward neural networks from the perspective of statistical pattern recognition, and is designed as a text, with over 100 exercises, to benefit anyone involved in the fields of neural computation and pattern recognition.

Classification and feature extraction of AVIRIS data

The processing of Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) data is discussed both in terms of feature extraction and classification. The recently proposed decision boundary feature

Radial Basis Functions, Multi-Variable Functional Interpolation and Adaptive Networks

Abstract : The relationship between 'learning' in adaptive layered networks and the fitting of data with high dimensional surfaces is discussed. This leads naturally to a picture of 'generalization

Classification of merged AVHRR and SMMR Arctic data with neural networks

A forward-feed back-propagation neural network is used to classify merged AVHRR and SMMR summer Arctic data. Four surface and eight cloud classes are identified. Partial memberships of each pixel to

Fast Learning in Networks of Locally-Tuned Processing Units

We propose a network architecture which uses a single internal layer of locally-tuned processing units to learn both classification tasks and real-valued function approximations (Moody and Darken

Parallel, self-organizing, hierarchical neural networks

A new neural-network architecture called the parallel, self-organizing, hierarchical neural network (PSHNN) is presented, which has many desirable properties, such as optimized system complexity, high classification accuracy, minimized learning and recall times, and truly parallel architectures in which all stages operate simultaneously without waiting for data from other stages during testing.

Pattern Recognition Principles

The present work gives an account of basic principles and available techniques for the analysis and design of pattern processing and recognition systems. Areas covered include decision functions,