Reconstruction from Feature Points of Face through Fuzzy C-Means Clustering Algorithm with Gabor Wavelets

FCM 군집화 알고리즘에 의한 얼굴의 특징점에서 Gabor 웨이브렛을 이용한 복원

  • 신영숙 (연세대학교 대학원 인지과학) ;
  • 이수용 (연세대학교 컴퓨터학과·산업시스템공학과) ;
  • 이일병 (연세대학교 컴퓨터학과·산업시스템공학과) ;
  • 정찬섭 (연세대학교 심리학과)
  • Published : 2000.06.01

Abstract

This paper reconstructs local region of a facial expression image from extracted feature points of facial expression image using FCM(Fuzzy C-Meang) clustering algorithm with Gabor wavelets. The feature extraction in a face is two steps. In the first step, we accomplish the edge extraction of main components of face using average value of 2-D Gabor wavelets coefficient histogram of image and in the next step, extract final feature points from the extracted edge information using FCM clustering algorithm. This study presents that the principal components of facial expression images can be reconstructed with only a few feature points extracted from FCM clustering algorithm. It can also be applied to objects recognition as well as facial expressions recognition.

본 논문은 FCM 군집화 알고리즘을 사용하여 표정영상에서 특징점들을 추출한 후 추출된 특징점으로부터 Gabor 웨이브렛들을 이용하여 표정영상의 국소영역을 복원한다. 얼굴의 특징점 추출은 두단계로 이루어진다. 1단계는 이차원 Gabor 웨이브렛 계수 히스토그램의 평균값을 적용하여 얼굴의 주요 요소성분들의 경계선을 추출한 후, 2단계에서는 추출된 경계선 정보로부터 FCM 군집화 알고리즘을 사용하여 얼굴의 주요 요소성분들의 최종적인 특징점들을 추출한다. 본 연구에서는 FCM 군집화 알고리즘을 이용하여 추출된 적은 수의 특징점들 만으로도 표정영상의 주요 요소들을 복원할 수 있음을 제시한다. 이것은 인간의 얼굴 표정인식 뿐만아니라 물체인식에도 적용되어질 수 있다.

Keywords

References

  1. Nebraska symposium on motivation Universal and cultural difference in facial expression of emotions Ekman, P.;Cloe, J. K.(ed.)
  2. IEICE Traction v.E 74 no.10 Recognition of facial expression from optical flow Mase, K.
  3. Journal of Cognitive Neuroscience v.3 no.1 Eigenfaces for recognition Turk, M. A.;Pentland, A. P.
  4. M.I.T. Media Group Technical Report no.194 Expression recognition using eigenfeatures Pentland, A.;Etocoff, N.;Starner, T.
  5. Proc. ICPR Real-time facial expression recognition based on features' positions and dimentions Sako, H.;Smith, A.
  6. '97 한국감성과학회 연차학술대회 논문집 기하학적인 특징 추출을 이용한 얼굴표정 인식 신영숙;이일병
  7. J. Opt. Soc. Amer. v.2 Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimentional visual cortical filters Daugman, J.
  8. J. Neurophys. v.58 An evaluation of the two-dimentional Gabor filter model of simple receptive fields in cat striate cortex Jones, J.;Parmer, L.
  9. Maximum entropy clusterings algorithms, System, Neural Nets, and Computing Technical Report Karayiannis, N. B.
  10. IEEE Transactions on Image Processing, in press Fuzzy vector quantization algorithms and their application in image compression Karayiannis, N. B.;Pai, P.
  11. Fuzzy mathematics in pattern classification Bezdek, J. C.
  12. IEEE Reans. Comput. v.42 no.3 Distortion invariant object recognition in the dynamic link architecture Lades, M.(et al.)
  13. Proc. of the IJCNN International Joint Conference on Neural Networks, San Diego, June pages II Size and distortion invariant object recognition by hierarchical graph matching Buhmann, J.;Lades, M.;von der Malsburg, C.
  14. Proc. of the ICANN96, Springer Verlag, Belin, Heidelberg, New York, Bochum, July Reconstruction from graphs labeled with responses of Gabor filters Potzsch, M.(et al.)