RUS  ENG
Full version
JOURNALS // Informatsionnye Tekhnologii i Vychslitel'nye Sistemy // Archive

Informatsionnye Tekhnologii i Vychslitel'nye Sistemy, 2015 Issue 2, Pages 40–49 (Mi itvs189)

PATTERN RECOGNITION

Image segmentation of color documents using color clustering

Yu. V. Vinogradovaa, D. P. Nikolaevb, D. G. Slugina

a Institute for Systems Analysis of Russian Academy of Sciences
b Institute for Information Transmission Problems of the Russian Academy of Sciences (Kharkevich Institute), Moscow

Abstract: In this paper, a novel method for segmentation of color images of documents for MRC compressions is proposed. The method is based on analyze of the color characteristic distribution of an image in the selected color space which gives good separability of the color clusters. The class of the considered documents is set and the mathematical model of the form of color distributions is determined. As the result, the three-dimensional color distribution is reduced to one-dimensional clustering of the histogram.

Keywords: image segmentation, color space, image compression, mixed raster content, MRC, PDF/A.



© Steklov Math. Inst. of RAS, 2024