DocumentCode
2545589
Title
Research on image compression algorithm based on Rectangle Segmentation and storage with sparse matrix
Author
Chen, Shengli ; Cheng, Xiaoxin ; Xu, Jiapin
Author_Institution
Sch. of Electron. & Inf. Eng., Sichuan Univ., SCU, Chengdu, China
fYear
2012
fDate
29-31 May 2012
Firstpage
1904
Lastpage
1908
Abstract
The Quarter-tree decomposition of image compression method characters with relative simplicity and fast calculation, however compression ratio is not very high. In order to overcome this flaw, one new segmentation method named the Rectangle Segmentation is proposed, in which adjacent pixel points satisfying consistency condition are viewed as the same image block. Also, without the restriction of square which abides to 2n, the image block can be rectangle which reduces the amount of block, and improves the compression ratio. Image compression ratio can be further augmented by combining the storage method of sparse matrix. Therefore, a new image compression algorithm is proposed named the Rectangle Segmentation and Sparse Matrix Storage(RSSMS) compression algorithm. Simulation results indicate that the compression ratios of images using the new algorithm is 25.19% higher than those using the Quarter-tree decomposition method.
Keywords
data compression; image coding; image segmentation; sparse matrices; RSSMS compression algorithm; compression ratio improvement; consistency condition; image blocks; image compression algorithm; image pixel points; quarter-tree decomposition; rectangle segmentation and sparse matrix storage compression algorithm; Compression algorithms; Gray-scale; Image coding; Image reconstruction; Image segmentation; PSNR; Sparse matrices; Image compression; Quarter-tree decomposition; rectangle segmentation; sparse matrix storage;
fLanguage
English
Publisher
ieee
Conference_Titel
Fuzzy Systems and Knowledge Discovery (FSKD), 2012 9th International Conference on
Conference_Location
Sichuan
Print_ISBN
978-1-4673-0025-4
Type
conf
DOI
10.1109/FSKD.2012.6233969
Filename
6233969
Link To Document