Genetic Algorithm Framework for Image Block Clustering

Jyh-Horng Jeng, Chih-Wen Wang


Principal Component Analysis (PCA) for image coding suffers the difficulties of computation complexity and projection errors. A typical method to solve this problem is the usage of clustering which partitions the image blocks into groups of smaller sizes. Moreover, the individuals belonging to the same group should exhibit the same visual properties such as edges and textures. In this paper, the genetic algorithm (GA) is adopted as a framework in the clustering process with visual properties imposed in the fitness function. Under such mechanism, the proposed method can effectively increase the retrieved quality and preserve the visual effects.


Genetic algorithm; K-means clustering; Principal component analysis

Full Text:



B. S. Everitt, S. Landaus, and M. Leese, Cluster Analysis, London, UK: Arnold, 4th edition, 2001.

P. De Lit, E. Falkenauer, and A. Delchambre, "Grouping genetic algorithms: an efficient method to solve the cell formation problem," Math. Comput. Simulat. 51, pp. 257-271, 2000.

J. Gertler and J. Cao, "PCA-based fault diagnosis in the presence of control and dynamic," AIChE J. 50, pp. 388–402, 2004.

J. Gertler and J. Cao, "Design of optimal residuals from partial principal component models for fault diagnosis in linear system," J. Process Control 15, pp. 585–603, 2005.

S. Costa "Fiori, Image compression using principal component neural networks," Image Vis. Comput. 19, pp. 649-668, 2001.

J. Gertler, W. Li, Y. Huang, and T. McAvoy, "Isolation enhanced principal component analysis," AIChE J. 45, pp. 323–334, 1999.

D. E. Goldberg, Genetic Algorithms in Search, Optimization, and Machine Learning. Addison Wesley, Massachusetts, 1989.

J. G. Hsieh, "A simple guide to machine learning and soft computing," (tutorial session speech). Proceeding of 14th International Conference on Intelligent System Applications to Power System, pp. 1-10, 2007.


  • There are currently no refbacks.

Copyright (c)