DocumentCode :
495361
Title :
Full Color Cartoon Image Lossless Compression Based on Region Segment
Author :
Zhe-lin, Li ; Qin-xiang, Xia ; Li-jun, Jiang ; Shi-zi, Wang
Author_Institution :
Sch. of Mech. & Automobile Eng., South China Univ. of Technol., Guangzhou, China
Volume :
6
fYear :
2009
fDate :
March 31 2009-April 2 2009
Firstpage :
545
Lastpage :
548
Abstract :
In accordance with the two-dimensional structural dependence of color cartoon image and the entropy coding theory in data compression, a method of color cartoon image lossless compression by eliminating the structural redundancy is presented. At first it is proved that the entropy of image can be reduced by way of removing the solid region. Then the method of removing the structural redundancy is discussed. The color image is divided into three kinds of region--background, solid region and gradient color region. In order to effectively recover data, the improved Freeman chain code and color seed are used to save the removed solid region information. We apply in sequence this method and the entropy coding method to eliminate the structural redundancy and the statistical redundancy for full color cartoon images. Experiments show that the proposed method results in reduction of bit rate by 25% compared to using directly universal methods of image lossless compression.
Keywords :
data compression; entropy codes; image coding; image colour analysis; image reconstruction; image segmentation; statistical analysis; Freeman chain code; background region; color seed; entropy coding theory; full color cartoon image lossless compression; gradient color region; image reconstruction; region segment; solid region; statistical redundancy; structural redundancy elimination; two-dimensional structural dependence; Automotive engineering; Computer science; Data engineering; Digital images; Entropy coding; Image coding; Image color analysis; Image segmentation; Pixel; Solids; Cartoon image; Entropy; Image Compression; Region Segment;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Science and Information Engineering, 2009 WRI World Congress on
Conference_Location :
Los Angeles, CA
Print_ISBN :
978-0-7695-3507-4
Type :
conf
DOI :
10.1109/CSIE.2009.672
Filename :
5170759
Link To Document :
بازگشت