Diffusing Vector Field of Gray-Scale Images for Image Segmentation
In this paper, a novel vector field transform is proposed for gray-scale images, which is based on the electrostatic analogy. By introducing the item of gray-scale difference into the repulsion vector between image points, the diffusing vector field is obtained by the transform. Then the primitive areas can be extracted based on the diffusing vector field with a proposed area-expanding method. A new image segmentation method is then presented by merging primitive areas. The experimental results prove the effectiveness of the proposed image segmentation method.
KeywordsVector Field Image Segmentation Test Image Image Point Background Area
- 1.YuJin Z (1999) Image engineering: image processing and analysis. TUP Press, Beijing, ChinaGoogle Scholar
- 3.Hurley DJ, Nixon MS, Carter JN (1999) Force field energy functionals for image feature extraction. In: Proceedings of the British machine vision conference, pp 604–613Google Scholar
- 5.Ahuja N, Chuang JH (1997) Shape representation using a generalized potential field model. IEEE T PAMI 19(2):169–176Google Scholar
- 6.Grogorishin T, Abdel-Hamid G, Yang YH (1996) Skeletonization: An electrostatic field-based approach. Pattern Anal Appl 1(3):163–177Google Scholar
- 8.Grant IS, Phillips WR (1990) Electromagnetism. 2nd ed. WileyGoogle Scholar