A novel color image quantization algorithm is developed. This new algorithm addresses the question of how to incorporate the principle of human visual perception to color variation sensitivity into color image quantization process. Color variation measure (CVM) is calculated first in CIE Lab color space. CVM is used to evaluate color variation and to coarsely segment the image. Considering both color variation and homogeneity of the image, the number of colors that should be used for each segmented region can be determined. Finally, CF-tree algorithm is applied to classify pixels into their corresponding palette colors. The quantized error of our proposed algorithm is small due to the combination of human visual perception and color variation. Experimental results reveal the superiority of the proposed approach in solving the color image quantization problem.

 Project Sponsors:

 Collaborators:

 Graduate Students:

 Yuchou Chang

1. Y. Chang and D.J. Lee, “Color Image Quantization Using Color Variation Measure”, The First IEEE Symposium on Computational Intelligence in Image and Signal Processing (CIISP 2007), Honolulu, Hawaii, USA, April 1-5, 2007.

(Click image to view paper poster)

Color Quantization Algorithm