The Journal of the Korea Contents Association (한국콘텐츠학회논문지)

The Korea Contents Association (한국콘텐츠학회)
권호 표지
DOI QR코드

DOI QR Code

Comparisons of Linear Feature Extraction Methods

선형적 특징추출 방법의 특성 비교

  • 오상훈 (목원대학교 정보통신공학과)
  • Published : 2009.04.28
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, feature extraction methods, which is one field of reducing dimensions of high-dimensional data, are empirically investigated. We selected the traditional PCA(Principal Component Analysis), ICA(Independent Component Analysis), NMF(Non-negative Matrix Factorization), and sNMF(Sparse NMF) for comparisons. ICA has a similar feature with the simple cell of V1. NMF implemented a "parts-based representation in the brain" and sNMF is a improved version of NMF. In order to visually investigate the extracted features, handwritten digits are handled. Also, the extracted features are used to train multi-layer perceptrons for recognition test. The characteristic of each feature extraction method will be useful when applying feature extraction methods to many real-world problems.

이 논문은 고차원의 데이터를 저 차원으로 줄이는 방법 중 하나인 특징추출에 대한 방법들의 특성을 비교한다. 비교대상 방법은 전통적인 PCA(Principal Component Analysis)방법과 시각피질의 특성을 보인다고 알려진 ICA(Independent Component Analysis), 국소기반인식을 구현한 NMF(Non-negative Matrix Factorization), 그리고 이의 성능을 개선한 sNMF(Sparse NMF)로 정하였다. 추출된 특징들의 특성을 시각적으로 확인하기 위하여 필기체 숫자 영상을 대상으로 특징추출을 수행하였으며, 인식기에 적용한 효과의 확인을 위하여 추출된 특징을 다층퍼셉트론에 학습시켜보았다. 각 방법의 특성을 비교한 결과는 응용하고자 하는 문제에서 어떤 특징을 추출하기 원하느냐에 따라 특징추출 방법을 선정할 때 유용할 것이다.

Keywords

References

  1. H. Kim and H. Park, "Sparse non-negative matrix factorizations via alternating non-negativity -constrained least squares for microarray data analysis," Bioinformatics, pp.1495-1502, 2007.
  2. Keinosuke Fukunaga, Introduction to Statistical Pattern Recognition, Elsevier, 1990.
  3. J. Ye, R. Janardan, Q. Li, and H. park, "Feature reduction via generalizaed uncrrelated linear discriment analysis," IEEE Trans. Knowledge and Data Engineering, Vol.18, pp.1312-1322, 2006. https://doi.org/10.1109/TKDE.2006.160
  4. D. D. Lee and H. S. Seung, "Learning the parts of objects by non-negative matrix factorization," Natute, Vol.401, pp.788-791, 1999. https://doi.org/10.1038/44565
  5. T.-W. Lee, M. Girolami, A. J. Bell, and T. J. Sejnowski, "A unfying information -theoretic framework for independent component analysis," Computers & Mathematics with Applications, Vol.31. pp.1-21, 2000.
  6. E. Doi, T. Inui, T-W. Lee, T. Wachtler, and T. J. Sejnowski, "Spatiochromatic receptive field properties derived from information-theoretic analyses of cone mosaic responses to natural scenes," Neural Computation, Vol.15, pp.397-417, 2003. https://doi.org/10.1162/089976603762552960
  7. S. Amari, "Natural gradient works efficiently in learning," Neural Computation, Vol.10, pp.251-276, 1998. https://doi.org/10.1162/089976698300017746
  8. A. J. Bell and T. J. Sejnowski, "The independent components of natural scenes are edge filters," Vision Res., Vol.37, pp.3327-3338, 1997. https://doi.org/10.1016/S0042-6989(97)00121-1
  9. J. B. Tenenbaum, V. de Silva, and J. C. Langford, "A global geometric framework for nonlinear dimensionality reduction," Science, Vol.290, No.22, pp.2319-2323, 2000. https://doi.org/10.1126/science.290.5500.2319
  10. S. T. Roweis and L. K. Saul, "Nonlinear dimensionality reduction by locally linear embedding," Science, Vol.290, No.22, pp.2323-2326, 2000. https://doi.org/10.1126/science.290.5500.2323
  11. J. J. Hull, "A database for handwritten text recognition research," IEEE Trans. Pat. Ana. Mach. Int., Vol.16, pp.550-554, 1994. https://doi.org/10.1109/34.291440
  12. S.-H. Oh, "Improving the error backpropagation algorithm with a modified error function," IEEE Trans. Neural Networks, Vol.8, pp.799-803, 1997. https://doi.org/10.1109/72.572117
  13. J.-M. Alonso and L. M. Martinez, "Functioanl connectivity between simple cells and complex cells in cat striate cortex," Nature Neuroscience, Vol.1, No.5, pp.395-403, 1998. https://doi.org/10.1038/1609