Hierarchical Tree of Image Derived by Diffusion Filtering
This paper aims to introduce a class of non-linear diffusion filterings based on deep structure analysis in scale space. In linear scale space, the trajectory of extrema is called stationary curves. This curves provides deep structure analysis and hierarchical expression of signals. The motion of extrema in linear scale space is controlled by a function of the higher derivatives of the signals. We introduce a non-linear diffusion filterings based on the absolute values of second derivative of signals.
KeywordsSingular Point Stationary Point Structure Tree Hessian Matrix Deep Structure
Unable to display preview. Download preview PDF.
- 1.Iijima, T.: Pattern Recognition, Corona-sha, Tokyo (1974) (in Japanese)Google Scholar
- 2.Zhao, N.-Y., Iijima, T.: Theory on the method of determination of view-point and field of vision during observation and measurement of figure. IEICE Japan, Trans. D. J68-D, 508–514 (1985) (in Japanese)Google Scholar
- 3.Zhao, N.-Y., Iijima, T.: A theory of feature extraction by the tree of stable view-point. IEICE Japan, Trans. D J68-D, 1125–1132 (1985) (in Japanese)Google Scholar
- 5.Kuijper, A.: The Deep Structure of Gaussian Scale Space Images, Ph.D. thesis, Utrecht University (2002)Google Scholar
- 14.Neelin, P.: McConnell Brain Imaging Centre, http://www.bic.mni.mcgill.ca/