Korean Journal of Remote Sensing (대한원격탐사학회지)

Korean Society of Remote Sensing (대한원격탐사학회)
권호 표지
DOI QR코드

DOI QR Code

Study on the Effect of Discrepancy of Training Sample Population in Neural Network Classification

  • Published : 2002.06.01
Download PDF
⟨ Previous Next ⟩

Abstract

Neural networks have been focused on as a robust classifier for the remotely sensed imagery due to its statistical independency and teaming ability. Also the artificial neural networks have been reported to be more tolerant to noise and missing data. However, unlike the conventional statistical classifiers which use the statistical parameters for the classification, a neural network classifier uses individual training sample in teaming stage. The training performance of a neural network is know to be very sensitive to the discrepancy of the number of the training samples of each class. In this paper, the effect of the population discrepancy of training samples of each class was analyzed with three layered feed forward network. And a method for reducing the effect was proposed and experimented with Landsat TM image. The results showed that the effect of the training sample size discrepancy should be carefully considered for faster and more accurate training of the network. Also, it was found that the proposed method which makes teaming rate as a function of the number of training samples in each class resulted in faster and more accurate training of the network.

Keywords

References

  1. IEEE Trans. Geosci. Remote Sensing v.28 Neural network approaches versus statistical methods in classification of multisource remote sensing Benediktsson, J. A.;Swain, P. H.;Erosy, O. K.
  2. Int. J. GIS. v.7 Artificial neural networks for landcover classfication and mapping Civco, D. L.
  3. Neural Networks v.5 Learning coefficient dependence on training set size Eaton, H. A. C.;Oliver, T. L.
  4. IEEE Trans. Geosci. Remote Sensing v.30 Classification of multispectral remote sensing data using a backpropagation neural network Heermann, P. D.;Khazenie, N.
  5. Photo. Eng. and Remote Sensing v.48 Artificial neural network classification using minimal training set: comparison to conventional supervised classification Hepner, G. F.;Logan, T.;Ritter, N.;Bryant, N.
  6. Korean Journal of Remote Sensing v.12 no.3 Trophic state mapping of Daechung reservoir by neural network classification of Landsat TM images Kim, K. E.;Kim, T. K.
  7. J. GIS Assoc. Korea. v.2 no.2 Study on application of neural network for unsupervised training of remote sensing data Kim, K. E.;Lee, T. S.
  8. Parallel Distributed Processig v.1 Learning internal representations by error propagation Rumelhart D. E.;Hinton G. E.;Williams, R. J.