A Graphic/Picture Classification Method Based on Spatial Characteristics of Colors

Abstract:

Article Preview

In order to get optimum color reproduction results on different color devices in color management, a proper rendering intent should be selected according to color characteristics of the image. In terms of differences on color characteristics, images could be classified into synthetic graphics and natural pictures. Different rendering intent should be applied on graphics and pictures. So graphic/picture automatic classification becomes a fundamental task of color management intellectualization. Characteristics on color distribution of a large number of images have been researched in our experiments. Then it is confirmed that the essential difference between graphic and picture is the characteristics on color distribution in the neighborhood of images rather than the number of colors or the volume of image gamut. Thus, the features which have distinct ability to show the differences could be used to build classification rules. In this paper, several mathematical features of image are extracted and selected by their classification performance. Based on these features, the discriminant analysis is adopted to build up discriminated functions. Finally, the accuracy of the functions has been tested and the precision is 96.75%.

Info:

Periodical:

Edited by:

Ouyang Yun, Xu Min, Yang Li and Liu Xunting

Pages:

31-35

Citation:

S. Y. Cai et al., "A Graphic/Picture Classification Method Based on Spatial Characteristics of Colors", Applied Mechanics and Materials, Vol. 262, pp. 31-35, 2013

Online since:

December 2012

Export:

Price:

$41.00

[1] Y.F. Wang, P. Zeng. Journal of Computer Application, Vol. 24 (2004), pp.70-72.

[2] Z.G. Fan U.S. Patent 0, 128, 369. (2003).

[3] F. Xie, L.T. Chen and H. Qiu. Application Research of Computers. Vol. 26 (2009), pp.2767-2770.

[4] X.Y. Ma. Microelectronics and Computer, Vol. 26 (2009), pp.250-252.

[5] S. Prabhakar, H. Cheng and Z.G. Fan U.S. Patent 0, 031, 268. (2002).

[6] M. Garbarino. Ms D Thesis. Stockholm. Royal Institute of Technology, (2008).

[7] Z. X Deng. Data Analyse Method and SAS System(Shanghai University of Finance and Econ- omic Press, Shanghai 2006).