Journal of the Institute of Convergence Signal Processing (융합신호처리학회논문지)

The Korea Institute of Convergence Signal Processing (한국융합신호처리학회)
권호 표지

Optimization of Max-Plus based Neural Networks using Genetic Algorithms

유전 알고리즘을 이용한 Max-Plus 기반의 뉴럴 네트워크 최적화

  • 한창욱 (동의대학교 전기공학과)
  • Received : 2012.10.05
  • Accepted : 2013.02.01
  • Published : 2013.01.30
Download PDF
⟨ Previous Next ⟩

Abstract

A hybrid genetic algorithm based learning method for the morphological neural networks (MNN) is proposed. The morphological neural networks are based on max-plus algebra, therefore, it is difficult to optimize the coefficients of MNN by the learning method with derivative operations. In order to solve the difficulty, a hybrid genetic algorithm based learning method to optimize the coefficients of MNN is used. Through the image compression/reconstruction experiment using test images extracted from standard image database(SIDBA), it is confirmed that the quality of the reconstructed images obtained by the proposed method is better than that obtained by the conventional neural networks.

본 논문에서는 하이브리드 유전 알고리즘을 이용한 morphological 뉴럴 네트워크 (MNN)의 최적화 방법을 제안하였다. MNN은 max-plus 연산을 기반으로 하고 있으므로 경사 학습법에 의한 파라미터 학습이 매우 어렵다. 이러한 문제를 해결하기 위해 하이브리드 유전 알고리즘을 이용하여 MNN의 파라미터들을 학습하였다. 제안된 방법의 유용성을 보이기 위해 SIDBA(standard image database) 표준영상에서 추출된 테스트 영상을 이용한 영상 압축/복원 실험을 수행하였고, 그 결과 제안된 방법에 의한 복원 영상이 합-곱 연산에 기반한 기존의 뉴럴 네트워크에 의한 복원영상보다 우수함을 알 수 있었다.

Keywords

References

  1. E. R. Dougherty and J. T. Astola(Eds.), "Nonlinear Filters for Image Processing," SPIE/IEEE Series on Imaging Science and Engineering, pp. 5-15, 1999.
  2. H. J. A. M. Heijmans, "Morphological Image Operators," Academic Press, pp. 3-20, Boston, 1994.
  3. R. Cuninghame-Green, "Minimax Algebra," Lecture Notes in Ecomomics and Mathematical Systems, Springer-Verlag, Vol. 166, pp. 4-30, New York, 1979.
  4. B. Heidergott et aI., "Max Plus at Work," Princeton Series in Applied Mathematics, pp. 13-42, 2006.
  5. J. L. Davidson and G. X. Ritter, "A Theory of Morphological Neural Networks," SPIE Digital Optical Computing II, Vol. 1215, pp. 378-388, 1990.
  6. G. X. Ritter, D. Li, and J. N. Wilson, "Image Algebra and Its Relationship to Neural Networks," SPIE Aerospace Pattern Recognition, Vol. 1098, pp. 90-101, 1989.
  7. R. Araujo, A. Oliveira, S. Soares, and S. Meira, "An evolutionary morphological approach for software development cost extimation," Neural Networks, Vol. 32, pp. 285-291, 2012. https://doi.org/10.1016/j.neunet.2012.02.040
  8. C. W. Han and J. I. Park, "SA-selection-based Genetic Algorithm for the Design of Fuzzy Controller," International Journal of Control, Automation, and Systems, Vol. 3, No. 2, pp. 236-243, 2005.
  9. J. Ortiz and C. Ocasio, "Analog hardware model for morphological neural networks," Int. Conf. Neural Networks and Computational Intelligence, pp. 40-44, 2003.
  10. D. E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, AddisonWesley, pp. 1-25, Reading, MA, 1989.
  11. S. Kirkpatrick, C. D. Gelatt Jr., and M. P. Vecchi, "Optimization by Simulated Annealing," Science, Vol. 220, No. 4598, pp. 671-680, 1983. https://doi.org/10.1126/science.220.4598.671