這個附註的目的是把二種設立初始權重的方法應用在Huang(2005)及其研究同仁所提出的 W-k- means運算法中。從彩色圖片分割的結果中，計算Liu 和 Yang (1994) 所提出的圖片分割求值函數，證明了我們所提出的方法比隨機產生權重的方法產生更好的分割結果。

The purpose of this note is to propose two attribute weight initialization methods in the W-k-means algorithm proposed by Huang et al. (2005). Results from color image segmentation illustrate the proposed methods produce better segmentation than the randomly generated weights according to Liu and Yang’s (1994) evaluation function.

[1] L. Boltzman, Neitere Studien uber das Warmegleichgewicht unter Gasmolekulen, K. Akad. Wiss. (Wein) Sitzb. 66 275, 1872.
[2] C. Cherry, On Human Communication, John Wiley & Sons, New York, 1957.
[3] R. Clausius, Abhaudlungen iiber die mechanische Warmetheorie Friedrich Vieweg, Braunschweig, 1864.
[4] J.Z. Huang, M.K. Ng, H. Rong, and Z. Li, Automated variable weighting in k-means type clustering. IEEE Trans. Pattern Anal. Machine Intelligence 27, 657-668, 2005.
[5] W.L. Hung, M.S. Yang, and D.H. Chen, Variation approaches to feature-weight selection and application to fuzzy clustering. IEEE International Conference on Fuzzy Systems, 276-280, 2008.
[6] S.S. Khan and A. Ahmad, Cluster center initialization algorithm for K-means clustering. Pattern Recognition Letters 25, 1293-1302, 2004.
[7] D.W. Kim, K.H. Lee, and D. Lee, A novel initialization for the fuzzy c-means algorithm for color clustering. Pattern Recognition Letters 25, 227-237, 2004.
[8] J. Liu and Y.H. Yang, Multiresolution color image segmentation technique, IEEE Trans. on Pattern Analysis and Machine Intelligence 16, 689-700, 1994.
[9] J. MacQueen, Some methods for classification and analysis of multivariate observation,Proc. Fifth Berkeley Symp. Math. Statistica and Probability, 281-297, 1967.
[10] C.E. Shannon, The mathematical theory of communication. Bell System Tech. J., 1948 and in Shannon and Weaver The Mathematical Theory of Communiction. University of Illinois, Urbana, 3-91, 1949.
[11] N. Wiener, Cybernetics Wiley, New York, 1949.