A Novel Image Edge Detection Using Fractal Compression
Image edges are the foundation of image texture and shape figure extraction. In this paper we propose a novel edge detection method based on the self-similarity of fractal compression. We point out that the mean-square-error distance (MSE) of fractal compression can be used to extract edge of fractal image effectively. The self-similarity coefficient between the local range block and the searching domain block is centered at the current pixel being processed, and near-center self-affine transform is applied in local searching process, finally a binary operator is used to threshold its magnitude and produce the edge map of the image. The results of experiments show that the proposed new algorithm for image edge detection is valid and effective. It also has good anti-noise performance.
Keywordsimage edge detection fractal coding MSE self-similarity
Unable to display preview. Download preview PDF.
- 1.Xu, J.H.: Image Process And Analysis, pp. 45–100. Academic Press, Beijing (1994)Google Scholar
- 4.Haralick, R.M.: Digital Step Edges from Zero Crossing of Second Directional Derivative. IEEE Trans. PAMI-6, 58–68 (1984)Google Scholar
- 5.Barnsley, M.F.: Iterated Function Systems and the Global Construction of Fractals. Proc. Roy. Soc. London, 243–275 (1985)Google Scholar
- 6.Jacquin, A.E.: A Novel Fractal Block-coding Technique for Digital Image. In: Proceedings of ICASSP IEEE International Conference on ASSP, vol. 4, pp. 34–42 (1990)Google Scholar
- 8.Shouji, C., Liming, Z.: Fractal And Image Compression. Shanghai Science And Technology Education Publishing House (1998)Google Scholar
- 11.Motallebi, P., Ogunbona, P.O.: Edge Image Description Using Fractal Interpolation. IEEE Tench-Speech and Image Technologies for Computing and Telecommunications, 553–556 (1997)Google Scholar
- 13.Qiong, L., Jun, G., Long, G., Huoming, D.: A Novel Edge Detection Method Based On Fractal Theory. In: IEEE int. Conf. Neural Network & Signal Processing, pp. 1105–1108 (2003)Google Scholar