Share Email Print
cover

Proceedings Paper

Statistical split and polynomial merge algorithm for image representation
Author(s): Seoung-Jun Oh; Keun-Heum Park
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Since polynomials fit the geometrical forms of images harmoniously and well represent slowly varying surfaces in images, there were many split and merge algorithms which used a polynomial function to represent each homogeneous region. Even though very low-bit rate can be achieved using their algorithms, it takes too much time for both split process and merging process. Furthermore, the splitted result is not quite well matched to HVS, either. In this paper, a new split and merge algorithm is designed. In this algorithm the split process uses a statistical hypothesis test called ShortCut method as a measurement of region homogeneity, and the merge process uses a polynomial function. The computation time for the split process can be significantly reduced using the new algorithm, and the new scheme reflects HVS more than any other scheme. To justify the algorithm proposed here, it is compared with other algorithms including Kunt's algorithm.

Paper Details

Date Published: 21 April 1995
PDF: 10 pages
Proc. SPIE 2501, Visual Communications and Image Processing '95, (21 April 1995); doi: 10.1117/12.206696
Show Author Affiliations
Seoung-Jun Oh, Kwangwoon Univ. (South Korea)
Keun-Heum Park, Kwangwoon Univ. (South Korea)


Published in SPIE Proceedings Vol. 2501:
Visual Communications and Image Processing '95
Lance T. Wu, Editor(s)

© SPIE. Terms of Use
Back to Top