DocumentCode :
2254227
Title :
Twice background determination algorithm for color document images
Author :
Tsai, Chung-Ming
Author_Institution :
Dept. of Comput. Sci., Taipei Municipal Univ. of Educ., Taipei, Taiwan
Volume :
5
fYear :
2010
fDate :
11-14 July 2010
Firstpage :
2692
Lastpage :
2696
Abstract :
Many color document images have been printed with different regions to highlight the content. Each region has different background color. Conventional document binarization methods cannot threshold these kind documents properly. Moreover, most of them are not efficient for processing color document with big image size. For example, the size of one page pdf file journal article is converted to 300 pixels per inch TIFF file format is 2481×3296. Conventional document binarization methods to threshold these big document images will take a lot of time. Thus, this paper proposes an efficient and effective two-stage background determination method for color document images to solve aforementioned problems. This novel method is very efficient because it only analysis the border pixels in the color document and the region images. Our experiments include various color document images. Experiment results show that the performance of the proposal method is better than other available thresholding methods in visual measurement.
Keywords :
document image processing; image colour analysis; image segmentation; aforementioned problems; color document images; conventional document binarization methods; thresholding methods; two-stage background determination method; Cybernetics; Image color analysis; Image segmentation; Machine learning; Optical character recognition software; Pixel; Proposals; Background determination; Color document; H-V scanning; Two-stage background determination; document binarization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
Type :
conf
DOI :
10.1109/ICMLC.2010.5580937
Filename :
5580937
Link To Document :
بازگشت